Categories
Admin Apache CentOS Linux Web Site Technologies

Major Headaches Migrating Apache from Ubuntu to CentOS

Intro
I’m changing servers from Ubuntu server to CentOS. On Ubuntu I just checked off LAMP and got my environment. In CentOS I’m doing it piece-by-piece. I don’t think my Ubuntu install is quite regular, either, as I bastardized it by adding environment variables in the Apache config file, a concept I borrowed from SLES! Turns out it is quite an ordeal to make a smooth transition. I will share all my pitfalls. I still don’t have it working, but I think I’m over the hump. [Update: now it is working, or 99% of it is working. It is a bit sluggish, however.]

The Details
I installed httpd on CentOS using yum. I also installed some php5 packages which I saw were recommended as well. First thing I noticed is that the directory structure for “httpd” as it seems to be known on CentOS, is dramatically different from “apache2” as it is known in Ubuntu. This example illustrates the point. In CentOS the main config file is

/etc/httpd/conf/httpd.conf

while in Ubuntu I had

/etc/apache2/apache2.conf

so I tarred up my /etc/apache2 files and had the thought “Let’s make this work on CentOS.” Ha. Easier said than done.

To remind, the content of /etc/apache2 is:

apache2.conf, conf.d, sites-enabled sites-available mods-enabled mods-available plus some stuff I probably added, including envvars, httpd.conf and ports.conf.

envvars contains environment variables which are subsequently referenced in the config files, like these:

export APACHE_RUN_USER=www-data
export APACHE_RUN_GROUP=www-data
export APACHE_PID_FILE=/var/run/apache2$SUFFIX.pid
export APACHE_RUN_DIR=/var/run/apache2$SUFFIX
export APACHE_LOCK_DIR=/var/lock/apache2$SUFFIX
# Only /var/log/apache2 is handled by /etc/logrotate.d/apache2.
export APACHE_LOG_DIR=/var/log/apache2$SUFFIX

First step? Well we have to hook httpd startup to our new directory somehow. I don’t recall this part so well. I think I tried this from the command line:

$ apachectl -d /etc/apache2 -f apache2.conf -k start

and it may be at that point that I got the MPM workers error. But I forget. I switched to using the service command and that particular error seemed to go away at some point. I don’t believe I needed to do anything special.

So I tried this edit to /etc/sysconfig/httpd (sparing you the failed attempts):

OPTIONS=”-d /etc/apache2 -f apache2.conf”

Now we try to launch and see what happens.

$ service httpd start

Starting httpd: httpd: Syntax error on line 203 of /etc/apache2/apache2.conf: Syntax error on line 1 of /etc/apache2/mods-enabled/alias.load: Cannot load /usr/lib/apache2/modules/mod_alias.so into server: /usr/lib/apache2/modules/mod_alias.so: cannot open shared object file: No such file or directory
[FAILED]

Fasten your seatbelts, put on your big-boy pants or whatever. We’re just getting warmed up.

Let’s look at mods-available/alias.load:

$ more alias.load

LoadModule alias_module /usr/lib/apache2/modules/mod_alias.so

Sure enough, there is not only no such file, there is not even such a directory as /usr/lib/apache2. And all the load files have references like that. Where did the httpd install put its modules anyways? Why in /etc/httpd/modules. So I made a command decision:

$ mkdir /usr/lib/apache2
$ cd !$
$ ln -s /etc/httpd/modules

So where does that put us? Here:

$ service httpd start

Starting httpd: httpd: Syntax error on line 203 of /etc/apache2/apache2.conf: Syntax error on line 1 of /etc/apache2/mods-enabled/ssl.load: Cannot load /usr/lib/apache2/modules/mod_ssl.so into server: /usr/lib/apache2/modules/mod_ssl.so: cannot open shared object file: No such file or directory
     [FAILED]

Not everyone will see this issue. I had used SSL for some testing in Ubuntu so I had that module enabled. my CentOS is a core image and did not come with an SSL module. So let’s get it.

$ yum search mod_ssl

shows the full module name to be mod_ssl.x86_64, so we install it with yum install.

How far did that get us? To here:

$ service httpd start

Starting httpd: httpd: bad user name ${APACHE_RUN_USER}
  [FAILED]

Ah, remember my environment variables from above? As I said I actually use them with lines such as:

User ${APACHE_RUN_USER}

in apache2.conf. But clearly the definitions of those environment variables is not getting passed along. I decide to see if this step might work. I append these two lines to /etc/sysconfig/httpd:

$ Read in our environment variables. Inspired by apache on SLES.
. /etc/apache2/envvars

Could any more go wrong? Sure. Lots! Then there’s this:

$ service httpd start

Starting httpd: httpd: bad user name www-data
      [FAILED]

Amongst all the other stark differences, ubuntu and CentOS use different users to run apache. Great! So I create a www-data user as userid 33, gid 33 because that’s how it was under ubuntu. but GID 33 is already taken in CentOS. It is backup. I decide I will never use it that way, and change the group name to www-data.

That brings us here. you see I have a lot of patience…

$ service httpd start

Starting httpd: Syntax error on line 218 of /etc/apache2/apache2.conf:
Invalid command 'LogFormat', perhaps misspelled or defined by a module not included in the server configuration
   [FAILED]

Now my line 218 looks pretty regular. It’s simply:

LogFormat "%v:%p %h %l %u %t \"%r\" %>s %O \"%{Referer}i\" \"%{User-Agent}i\"" vhost_combined

I then realized something interesting. The modules built in to httpd on centOS and apache2 are different. apache2 seems to have some modules built in for logging:

$ apache2 -l

Compiled in modules:
  core.c
  mod_log_config.c
  mod_logio.c
  prefork.c
  http_core.c
  mod_so.c

whereas httpd does not:

$ httpd -l

Compiled in modules:
  core.c
  prefork.c
  http_core.c
  mod_so.c

So I made an empty log_config.conf and a log_config.load in mods-available that reads:

LoadModule log_config_module /usr/lib/apache2/modules/mod_log_config.so

I got the correct names by looking at the apache web site documenttion on that module. And i linked those two files up in the mods-available diretory:

$ cd mods-enabled
$ ln -s ../mods-available/log_config.conf
$ ln -s ../mods-available/log_config.load

Next error, please! Certainly. It is:

$ service httpd start

Starting httpd: Syntax error on line 218 of /etc/apache2/apache2.conf:
Unrecognized LogFormat directive %O
  [FAILED]

where line 218 is as recorded above. Well, some searches showed that you need the logio module. Note that it is also compiled into to apache2, but missing from httpd. So I did a similar thing with defining the necessary mods-{available,enabled} files. logio.load reads:

LoadModule logio_module /usr/lib/apache2/modules/mod_logio.so

The next?

$ service httpd start

Starting httpd: (2)No such file or directory: httpd: could not open error log file /var/log/apache2/error.log.
Unable to open logs
   [FAILED]

Oops. Didn’t make that directory. Naturally httpd and apache2 use different directories for logging. What else could you expect?

Now we’re down to this minimalist error:

$ service httpd start

Starting httpd:     [FAILED]

The error log contained this line:

[Mon Mar 19 14:11:14 2012] [error] (2)No such file or directory: Cannot create SSLMutex with file `/var/run/apache2/ssl_mutex'

After puzzling some over this what I eventually noticed is that my environment has references to directories which I haven’t defined yet:

export APACHE_RUN_DIR=/var/run/apache2$SUFFIX
export APACHE_LOCK_DIR=/var/lock/apache2$SUFFIX

So I created them.

And now I get:

$ service httpd start

Starting httpd:           [  OK  ]

But all is still not well. I cannot stop it the proper way. Trying to read its status goes like this:

$ service httpd status

httpd dead but subsys locked

I looked this one up. Killed off processes and semaphores as recommended with ipcs -s (see this link), etc. But since my case is different, I also did something different. I modified my /etc/init.d/httpd file:

#pidfile=${PIDFILE-/var/run/httpd/httpd.pid}
pidfile=${PIDFILE-/var/run/apache2.pid}
#lockfile=${LOCKFILE-/var/lock/subsys/httpd}
lockfile=${LOCKFILE-/var/lock/subsys/apache2}

Believe it or not, this worked. I can now run service httpd status and service httpd stop. To prove it:

$ service httpd status

httpd (pid  30366) is running...

Another Error Crops Up
I eventually noticed another problem with the web site. My trajectory page was not working. Upon investigation I found this comment in my main apache error log (not my virtual server error log, which I still don’t understand):

sh: /home/drj/traj/traj4.pl: Permission denied

This had to be a result of my call-out to a perl program from a php program:

...
$data = exec('/home/drj/traj/traj4.pl'.' '.$escargs);
...

But what’s so special about that call? Worked fine on Ubuntu, and I did a directory listing to show the file was really there. Well, here’s the thing, that file is under my home directory and guess what? When you crate your users in Ubuntu the home directory permissions are set to group and others read. Not in CentOS! A listing of /home looks kind of like this:

/home$ ll

total 12
drwx------ 2 drj   drj     4096 Mar 19 15:26 drj/
...

I set the permissions for all to read:

$ sudo chmod g+rx,o+rx drj

and I was good to go. The program began to work.

May 2013 Update
I was asked how all this survived after a yum update. Answer: pretty well, but not perfectly. The daemon was fine. And what miseld me is that it started fine. But then a couple days later I was looking at my access log and realized…it wasn’t there! Nor the errors log. Well, actually, the default access and error logs were there, but not for my virtual servers.

I soon realized that

$ service httpd status

produced

httpd dead but subsys locked

Well, who reads or remembers their own posts from a year ago? I totally forgot I had already dealt with this once, and my own post didn’t show up in my DDG search. Anywho, I stepped on the same rake twice. Being less patient this time around, probably because I am one year older, I simply altered the /etc/init.d/httpd file (looks like it had been changed by the update) thusly:

#pidfile=${PIDFILE-/var/run/httpd/httpd.pid}
#lockfile=${LOCKFILE-/var/lock/subsys/httpd}
# try as an experiment - DrJ 5/3/13
pidfile=/var/run/apache2.pid
lockfile=/var/lock/apache2/accept.lock

and I made sure I had a /var/lock/apache2 directory. This worked.

I chose a lock file with that particular name because I noticed this in my /etc/apache2/apache2.conf:

LockFile ${APACHE_LOCK_DIR}/accept.lock

To clean things out as I was working and re-working this problem since I couldn’t run

$ service httpd stop

I ran instead:

$ pkill -9 -f sbin/httpd

and I removed /var/run/apache2.pid.

Now, once again, I can get a status on my httpd service and restart works as well and my access and error logs are being written.

Conclusion
This conversion exercise turned out to be quite a teaching lesson and even after all this more remains. After the mysql migration I find the performance to be sub-par – about twice as slow as it was on Ubuntu.

Four months later, CentOS has not crashed once on me. Contrast that with Ubuntu freezing every two weeks or so. I optimized MySQL to cache some data and performance is adequate. I also have since learned about bitnami, which is kind of a stack for all the stuff I was using. Check out bitnami.org.

Categories
Admin Hosting Service Linux

Hosting: You Really Can’t beat Amazon Web Services EC2

Intro
You want to have your own server hosted by a service provider that’s going to take care of the hard stuff – uninterruptible power, fast pipe to the Internet, backups? That’s what I wanted. In addition I didn’t want to worry about actual, messy hardware. Give me a virtual server any day of the week. I am no hosting expert, but I have some experience and I’d like to share that.

The Details
I’d say chances are about even whether you’d think of Amazon Web Services for the above scenario. I’d argue that Amazon is actually the most competitive service out there and should be at the top of any short list, but the situation wasn’t always so even as recently as February of this year.

You see, Amazon markets itself a bit differently. They are an IaaS (infrastructure as a service) provider. I don’t know who their top competition really is, but AWS (Amazon Web Service) is viewed as both visionary and able to execute by Gartner from a recent report. My personal experience over the last 12 months backs that up. My main point, however, is that hosting a server is a subset of IaaS. I assume that if you want your own server where you get root access, you have the skill set (aided by the vast resources on the Internet including blogs like mine) to install a web server, database server, programming environment, application engines or whatever you want to do with it. You don’t need the AWS utility computing model per se, just a reliable 24×7 server, right? That’s my situation.

I was actually looking to move to “regular” hosting provider, but it turns out to have been a really great time to look around. Some background. I’m currently running such an environment running Ubuntu server 10.10 as a free-tier micro instance. I’ve enjoyed it a lot except one thing. From time to time my server freezes. At least once a month since December. I have no idea why. Knowing that my free tier would be up anyways this month I asked my computer scientist friend “Niz” for a good OS to run a web server and he said CentOS is what I want. It’s basically Redhat Enterprise Linux except you don’t pay Redhat for support.

I looked at traditional hosting providers GoDaddy and Rackspace and 1and1 a bit. I ran the numbers and saw that GoDaddy, with whom I already host my DNS domains, was by far the cost leader. They were also offering CentOS v 5.6 I think RackSpace also had a CentOS offering. I spoke with a couple providers in my own state. I reasoned I wuold keep my business local if the price was within 25% of other offers I could find.

Then, and here’s one of the cool things about IaaS, I fired up a CentOS image at Amazon Elastic Compute Cloud. With utility computing I pay only by the hour so I can experiment cheaply, which I did. Niz said run v 5.6 because all the bugs have been worked out. He hosts with another provider so he knows a thing or two about this topic and many other topics besides. I asked him what version he runs. 5.6. So I fired it up. But you know, it just felt like a giant step backwards through an open source timeline. I mean Perl v 5.8.8 vs Ubuntu’s 5.10.1. Now mind you by this time my version of Ubuntu is itself a year old. Apache version 2.2.3 and kernel version 2.6.18 versus 2.2.16 and 2.6.35. Just plain old. Though he said support would be available for fantastical amount of time, I decided to chuck that image.

Just as I was thinking about all these things Amazon made a really important announcement: prices to be lowered. All of a sudden they were competitive when viewed as a pure hosting provider, never mind all the other features they bring to bear.

I decided I wanted more memory than the 700 MB available to a micro image, and more storage than the 8 GB that tier gives. So a “small” image was the next step up, at 1.7 GB of memory and 160 GB disk space. But then I noticed a quirky thing – the small images only come in 32-bit, not 64-bit unlike all the other tiers. I am so used to 64-bit by now that I don’t trust 32-bit. I want to run what a lot of other people are running to know that the issues have been worked out.

Then another wonderful thing happened – Amazon announced support for 64-bit OSes in their small tier! What timing.

The Comparison Results
AWS lowered their prices by about 35%, a really substantial amount. I am willing to commit up front for an extended hosting because I expect to be in this for the long haul. Frankly, I love having my own server! So I committed to three years small tier, heavy usage after doing the math in order to get the best deal for a 24×7 server. It’s $300 $96 up front and about $0.012$0.027/hour for one instance hour. So that’s about $18 $22/month over three years. Reasonable, I would say. For some reason my earlier calculations had it coming out cheaper. These numbers are as of September, 2013. I was prepared to use GoDaddy which I think is $24/month for a two-year commitment. My finding was that RackSpace and 1and1 were more expensive in turn than GoDaddy. I have no idea how AWS did what they did on pricing. It’s kind of amazing. My local providers? One came in at six times the cost of GoDaddy(!), the other about $55/month. Too bad for them. But I am excited about my new server. I guess it’s a sort of master of my own destiny type of thing that appeals to my independent spirit. Don’t tell Amazon, but really I think they could have easily justified charging a small premium for their hosting, given all the other convenient infrastructure services that are there, ready to be dialed up, say, like a load balancer, snapshots, additional IPs, etc. And there are literally 8000 images to choose from when you are deciding what image (OS) to run. That alone speaks volumes about the choices you have available.

What I’m up to
I installed CentOS 6.0 core image. It feels fresher. It’s based on RedHat 6.0 It’s got Perl v. 5.10.1, kernel 2.6.32, and, once you install it, Apache v 2.2.15. It only came with about 300 packages installed, which is kind of nice, not the usual 1000+ bloated deal I am more used to. And it seems fast, too. Now whether or not it will prove to be stable is an entirely different question and only time will tell. I’m optimistic. But if not, I’ll chuck it and find something else. I’ve got my data on a separate volume anyways which will persist regardless of what image I choose – another nice plus of Amazon’s utility computing model.

A Quick Tip About Additional Volumes
With my micro instance it occupied a full 8 GB so I didn’t have a care about additional disk volumes. On the other hand, my CentOS 6.0 core image is a lean 6 GB. If I’m entitled to 160 GB as part of what I’m paying for, how do I get the access to the remaining 154 GB? I guess you create a volume. Using the Admin GUI is easiest. OK, so you have your volunme, how does your instance see it? It’s not too obvious from their documentation but in CentOS my extra volume is

/dev/xvdj

I mounted that a formatted it as an ext4 device as per their instructions. It didn’t take that long. I put in a line in /etc/fstab like this:

/dev/xvdj /mnt/vol ext4 defaults 1 2

Now I’m good to go! It gets mounted after reboot.

Dec, 2016 update
Amazon has announced Lightsail to better compete with GoDaddy and their ilk. Plans start as low as $5 a month. For $10 a month you get a static IP, 1 GB RAM, 20 GB SSD storage I think and ssh access. So I hope that means root access. Oh, plus a pre-configured WordPress software.

Conclusion
Amazon EC2 rocks. They could have charged a premium but instead they are the cheapest offering out there according to my informal survey. The richness of their service offerings is awesome. I didn’t mention that you can mount the entire data set of the human genome, or all the facts of the world which have been assembled in freebase.org. How cool is that?

Categories
Admin Linux SLES

How to Get By Without unix2dos in SLES

Intro
As a Unix old-timer looking at the latest releases, I only have observed one tendency – that of ever-increasing numbers of commands, always additive – until now. A command I considered useful (well, basically any command I have ever used I consider useful) has gone AWOL in Suse Linux Enterprise Server (SLES for short): unix2dos.

Why You Need It
These days you need it more than ever. What with sftp being used in place of ftp, your transferred text files will come over from a SLES server to your PC in binary mode, preserving the Linux-style way of declaring a new line with the newline character, “\n”. Bring that file onto your PC and look at it in Notepad and you’ll get one long line because Windows requires more to indicate a new line. Windows OS’s like Windows 7 require a carriage return + newline, i.e., “\r\n”.

Who You Going to Call
I spoke with some experts so I cannot take credit for finding this out personally. Long story short things evolved and there is a more sophisticated command available that does this sort of thing and much else. That’s recode.

But I don’t think I’ll ever use recode for anything else so I decided to re-create a unix2dos command using recode in a tiny shell script:

#!/bin/sh
# inspired by http://yourlinuxguy.com/?p=232 and the fact that they took away this useful command
# 3/6/12
recode latin1..ibmpc $*

You call it like this:

> unix2dos file

and it overwrites the file and converts it to the format Windows expects.

My other expert contact says I could find the old unix2dos in OpenSuse but I decided not to go that route.

Of course to convert in the other direction you have dos2unix which for some reason wasn’t removed from the distro. Strange, huh?

How to See That It Worked
I use

> od -c file|more

to look at the ascii characters in a text file. It also shows the newline and carriage return characters with a \n and \r respectively This is a good command to know because it is also a “safe” way to look at a binary file. By safe I mean it won’t try to print out 8-bit characters that will permanently mess your terminal settings!

2017 update
I finally needed this utility again after five years. My program doesn’t work on CentOS. – No recode, whatever that was. However, the one-liner provided in the comments worked just fine for me.

Conclusion
We can rest easy and send text files back-and-forth between a PC and a SLES server with the help of this unix2dos script we developed.

Interestingly, RedHat’s RHEL has kept unix2dos in its disrtibution. Good for them. In ubuntu Linux unix2dos also seems decidedly missing.

Categories
Admin Linux

Common Problems Installing Cognos Gateway on Linux

Updated for a 2018 Cognos 11 install
with 2013 updates for Cognos 10 installation

Intro
I tried to take a shortcut and get a 2nd Cognos gateway up and running by copying files, etc. rather than a proper install. At one time or another I feel I must have encountered just about every problem conceivable. I didn’t take great, systematic notes, but I’d like to mention some highlights while it is still fresh in my memory!

The Details
Note that I have a working gateway server running on the same version of Linux, SLES 11 SP1. So I thought I’d be clever and just copy all the files below /opt/cognos8 from the working server.

First Rookie Mistake
Let’s call our COGNOS_ROOT /opt/cognos8 for convenience.
Cognos 10 note: /opt/cognos10 would be a more sensible installation directory!

So you’re following along in the documentation and dutifully looking for /opt/cognos8/bin/cogconfig.sh, and not finding it? Me, neither. So I cleverly borrowed it from a working solaris installation. It’s all Java, right, no OS dependencies, what can go wrong? Ha, ha. You try:

./cogconfig.sh
and get:

Using /usr/lib64/jvm/jre/bin/java
The java class is not found:  CRConfig

Long story short. Give up. Without telling anyone they moved it to /opt/cognos8/bin64. That’s assuming you’re on a 64-bit system like most of us are.

OK. Now you run it from the …bin64 directory, expecting better results, only to perhaps get something like:

./cogconfig.sh

Unable to locate a JRE. Please specify a valid JAVA_HOME environment variable.

Long story short, java-1_4_2-ibm (java-1_6_0-ibm if installing a Cognos 10 gateway) is a good Java environment to install for Cognos Gateway. At least it is on SLES Linux. So you install that and set up environment variables like these:

export JAVA_BINDIR=/usr/lib64/jvm/jre/bin
export JAVA_HOME=/usr/lib64/jvm/jre
export JAVA_ROOT=/usr/lib64/jvm/jre

Now you’re cooking. Run it yet again. You’re smart and know to set up your DISPLAY environment to a valid XServer you have access to. But even if the X application actually does launch and run (you may need some Motif or additional X packages, possibly even from the SDK DVD – see appendix A), if you try to export the configuration you’ll get an error like this:

java.lang.ClassNotFoundException: org.bouncycastle134.jce.provider.BouncyCastleProvider

Cognos 10 note: I did not have this class missing in my Cognos 10 installation. Yeah!

Yes, you are missing the infamous bouncycastleprovider! This stuff is too good to make up, right? It’s a jar file that’s somewhere in the Cognos Gateway distribution, bcprov-jdk14-134.jar. In my case I need to put it here:

/etc/alternatives/jre/lib/ext

With that in place run it yet again. Now you may be unable to export the configuration with this error:

CAM-CRP-1057 Unable to generate the machine specific symmetric key.

Does it ever end? Yes!

You may have old values of keys and what-not cryptography stuff from your copy of the other system. So you remove these directories and all their contents:

/opt/cognos8/{encryptkeypair,signkeypair}

And I even saw the following error:

02/03/2012,11:26:56,Err,com.cognos.crconfig.data.DataManagerException: CAM-CRP-1132 An error occurred while attempting to request a certificate from the Certificate Authority service. Unable to connect to the Certificate Authority service. Ensure that the Content Manager computer is configured and that the Cognos 8 services on it are currently running. Reason: java.net.ConnectException: Connection refused, com.cognos.crconfig.data.DataManager.generateCryptoKeys(DataManager.java:2730)

I think it comes about if you save the default config without editing it and putting in a valid dispatcher URI, but I forget.

The main point towards the end was to start with a clean config by a:

cd /opt/cognos8/configuration;cp cogstartup.xml{.new,}

, making sure there is no encryptkeypair and signkeypair directories, launching …bin64/cogconfig.sh, working with the GUI to define the dispatcher URIs to your working, running Cognos dispatcher, exporting it,

(Let me take a breath here. If that export succeeds, you’re home.)

and finally saving it, which also generates the system-specific keys.

That’s it! A bunch of green check marks are your reward. Hopefully.

Conclusion
In the end you will see that this “cheap method” of installing Cognos Gateway worked. We had a few bumps along the road, but we worked through them all. Now that we’ve seen just about every conceivable problem we have a treasure trove of documented errors and fixes should we ever find ourselves in this situation again.

There is one more Cognos Gateway problem we resolved, by the way, that was previously documented here.

Appendix A – Cognos 10 note
Yes, I referred to this document in my own installation of Cognos version 10 gateway component. The problems are very similar, and this was a big help, if I say so myself.

I notice I write a tight narrative. I have lots of tangential thoughts, but to list them all as I think of them would destroy the flow of the narrative. In this case I wanted to expand on the openmotif packages.

I got a missing libXm.so.4 message when launching issetup the first time. I determined this came from an openmotif package from my previous successful installation on another server. My new server had limited repositories.

> zypper search openmotif

produced these results:

 
S | Name                   | Summary                    | Type
--+------------------------+----------------------------+-----------
  | openmotif21-demos      | Open Motif 2.2.4 Libraries | package
  | openmotif21-libs       | Open Motif 2.2.4 Libraries | package
  | openmotif21-libs       | Open Motif 2.2.4 Libraries | srcpackage
  | openmotif21-libs-32bit | Open Motif 2.2.4 Libraries | package
  | openmotif22-libs       | Open Motif 2.2.4 Libraries | package
  | openmotif22-libs       | Open Motif 2.2.4 Libraries | srcpackage
  | openmotif22-libs-32bit | Open Motif 2.2.4 Libraries | package

Well, I tried to install first openmotif21-libs-32bit then openmotif22-libs-32bit, but neither gave me the right version of libXm.so! I had versions 2, 3 and 6! So I simply did one of these numbers:

> cd /usr/lib; ln -s libXm.so.3.0.3 libXm.so.4

and, to my surprise, it worked!

More Errors Documented for completeness’ sake

At the risk of making this blog post a total mess, I’ll include a few more errors I encountered during the upgrade. Who knows who might find this useful.

Generating the cryptographic keys is always a hold-your-breath-and-pray operation. I had my upgrade files in place in a new install directory, /opt/cognos10. I ran bin64/cogconfig.sh like usual. It was suggested I could save the configuration even though the application gateway wasn’t running, so I tried that. No dice.

The cryptographic information cannot be encrypted.

Fine. So probably the app server needs to be running before we save the config, right? So they got it running. I tried to save the config. Same error. The details were as follows:

[ ERROR ]
CAM-CRP-1315 Current configuration points to a different Trust Domain than originally configured.
 
[ ERROR ] 
The cryptography information was not generated.

The remedy? Close the configuration and completely remove these directories beneath the /opt/cognos10/configuration directory:

– encryptkeypair
– signkeypair
– csk (actually I didn’t have this one. But I guess it should be removed if present)

I held my breath, re-ran cogconfig and saved. This time it worked!

I also had an error with my Java version:

./cogconfig.sh
Using /usr/lib64/jvm/jre/bin/java
The java class could not be loaded. java.lang.UnsupportedClassVersionError: (CRConfig) bad major version at offset=6
/usr/lib64/jvm/jre/bin/java -version

showed

java version "1.4.2"
Java(TM) 2 Runtime Environment, Standard Edition (build 2.3)
IBM J9 VM (build 2.3, J2RE 1.4.2 IBM J9 2.3 Linux amd64-64 j9vmxa64142ifx-20110628 (JIT enabled)
J9VM - 20110627_85693_LHdSMr
JIT  - 20090210_1447ifx5_r8
GC   - 200902_24)

I installed a newer Java:

zypper install  java-1_6_0-ibm

and got past this error.

April 20123 update
Just when you thought every possible error was covered, you encounter a new one. Cognos Mobile isn’t working so well on actual mobile devices so they wanted to try a Fixpack from IBM. No problem, right? They gave me

up_cogmob_linuxi38664h_10.2.1102.33_ml.rar

and I set to work. I don’t particularly like rar files for Linux, but I figured out there is an unrar command:

$ unrar e up_*rar

But after setting up my DISPLAY environment variable I get this new error running ./issetup:

X Error of failed request:  BadDrawable (invalid Pixmap or Window parameter)
  Major opcode of failed request:  14 (X_GetGeometry)
  Resource id in failed request:  0x2
  Serial number of failed request:  257
  Current serial number in output stream:  257
IDS_MSG_PREFIXIDS_COPYRIGHT_LOGOIDS_MSG_PREFIXIDS_MSG_READ_ARCHIVE

The solution? They downloaded a tar.gz version of the Fixpack. I unpacked that and had absolutely no problems with issetup! The really strange thing is that in both issetup are identical files. I use cksum to do a quick compare. Even setup.csp are identical files. I did an strace -f of the two cases but the salient difference didn’t pop out at me. The files present in the tar.gz seem to be fewer in number.

Another random error you will encounter sooner or later

You are doing a Save in cogconfig and you get:

13/05/2013,17:39:05,Err,CAM-CRP-1132 An error occurred while attempting to request a certificate from the Certificate Authority service. Unable to connect to the Certificate Authority service. Ensure that the Content Manager computer is configured and that the IBM Cognos services on it are currently running. Reason: java.net.ConnectException: Connection refused, com.cognos.crconfig.data.crypto.ConfiguringSession.configure(ConfiguringSession.java:35)com.cognos.crconfig.data.DataManager.generateCryptoKeys(DataManager.java:3037)com.cognos.crconfig.data.DataManager$4.run(DataManager.java:4169)com.cognos.crconfig.data.CnfgActionEngine$CnfgActionThread.run(CnfgActionEngine.java:394)com.cognos.crconfig.data.crypto.ConfiguringSession.configure(ConfiguringSession.java:35)com.cognos.crconfig.data.DataManager.generateCryptoKeys(DataManager.java:3037)com.cognos.crconfig.data.DataManager$4.run(DataManager.java:4169)com.cognos.crconfig.data.CnfgActionEngine$CnfgActionThread.run(CnfgActionEngine.java:394)com.cognos.crconfig.data.crypto.ConfiguringSession.configure(ConfiguringSession.java:35)com.cognos.crconfig.data.DataManager.generateCryptoKeys(DataManager.java:3037)com.cognos.crconfig.data.DataManager$4.run(DataManager.java:4169)com.cognos.crconfig.data.CnfgActionEngine$CnfgActionThread.run(CnfgActionEngine.java:394)

This looks scary but has an easy fix. You aren’t communicating with the app server. Probably their dispatcher services are down. Bring them up and it should work fine – it did for me. This is assuming of course that you have your dispatcher URLs set up correctly.

I cloned my Cognos web gateway and got this error
I waited for a few weeks to examine the clone. I ran

$ ./cogconfig.sh

and got this error:

16/05/2013,15:57:35,Err,CAM-CRP-1280 An error occurred while trying to decrypt using the system protection key. Reason: javax.crypto.IllegalBlockSizeException: Input length (with padding) not multiple of 16 bytes

Umm. I don’t have the solution yet. One thing is most highly suspect: in the meatime we re-generated the keys on the production web gateway. So I am hoping that is all we need to do here as well.

Resolved. Here is the process I followed – a sort of colonic for Cognos:

$ cd /opt/cognos10/configuration; rm csk/* signkeypair/* encryptkeypair/* cogstartup.xml
$ cd ../bin64; ./cogconfig.sh

Then in the GUI I re-defined the app servers in the dispatcher URI portion of the environment.
Then did a Save.
Worked like a champ – four green check marks.

cogconfig hangs
This happened to me on an older server. The IBM Cognos Configuration screen displays but it’s supposed to exit so you can get to the part where you edit the configuration and it never does.

Currently no known solution.

June 2018 update
Cognos 11 install problem

The Cognos 11 install was going pretty well. Until it came time to launch cogconfig. That generated this error:

cognos10:/web/cognos11/bin64> ./cogconfig.sh

Using /usr/lib64/jvm/jre/bin/java
Exception in thread "main" java.lang.UnsupportedClassVersionError: JVMCFRE003 bad major version; class=com/cognos/accman/jcam/crypto/CAMCryptoException, offset=6
        at java.lang.ClassLoader.defineClass(ClassLoader.java:286)
        at java.security.SecureClassLoader.defineClass(SecureClassLoader.java:74)
        at java.net.URLClassLoader.defineClass(URLClassLoader.java:538)
        at java.net.URLClassLoader.defineClass(URLClassLoader.java:449)
        at java.net.URLClassLoader.access$300(URLClassLoader.java:77)
        at java.net.URLClassLoader$ClassFinder.run(URLClassLoader.java:1041)
        at java.security.AccessController.doPrivileged(AccessController.java:448)
        at java.net.URLClassLoader.findClass(URLClassLoader.java:427)
        at java.lang.ClassLoader.loadClass(ClassLoader.java:676)
        at sun.misc.Launcher$AppClassLoader.loadClass(Launcher.java:358)
        at java.lang.ClassLoader.loadClass(ClassLoader.java:642)
        at java.lang.J9VMInternals.verifyImpl(Native Method)
        at java.lang.J9VMInternals.verify(J9VMInternals.java:73)
        at java.lang.J9VMInternals.initialize(J9VMInternals.java:133)
        at com.cognos.cclcfgapi.CCLConfigurationFactory.getInstance(CCLConfigurationFactory.java:59)
        at com.cognos.crconfig.CnfgPreferences.<init>(CnfgPreferences.java:51)
        at com.cognos.crconfig.CnfgPreferences.<clinit>(CnfgPreferences.java:36)
        at java.lang.J9VMInternals.initializeImpl(Native Method)
        at java.lang.J9VMInternals.initialize(J9VMInternals.java:199)
        at CRConfig.main(CRConfig.java:144)

Note my system java version is woefully out-of-date:

$ /usr/lib64/jvm/jre/bin/java ‐version

java version "1.6.0"
Java(TM) SE Runtime Environment (build pxa6460sr16fp15-20151106_01(SR16 FP15))
IBM J9 VM (build 2.4, JRE 1.6.0 IBM J9 2.4 Linux amd64-64 jvmxa6460sr16fp15-20151020_272943 (JIT enabled, AOT enabled)
J9VM - 20151020_272943
JIT  - r9_20151019_103450
GC   - GA24_Java6_SR16_20151020_1627_B272943)
JCL  - 20151105_01

whereas the Cognos-supplied Java is two versions ahead:
cognos10:/web/cognos11> ./jre/bin/java ‐version

java version "1.8.0"
Java(TM) SE Runtime Environment (build pxa6480sr4fp10-20170727_01(SR4 FP10))
IBM J9 VM (build 2.8, JRE 1.8.0 Linux amd64-64 Compressed References 20170722_357405 (JIT enabled, AOT enabled)
J9VM - R28_20170722_0201_B357405
JIT  - tr.r14.java_20170722_357405
GC   - R28_20170722_0201_B357405_CMPRSS
J9CL - 20170722_357405)
JCL - 20170726_01 based on Oracle jdk8u144-b01

Instead of the previous approach which involved upgrading the system Java, I decided to just try the Java version Cognos itself had installed. In the following commands note that my installation directory was /web/cognos11.

$ cd /web/cognos11; export JAVA_HOME=`pwd`/jre
$ ./cogconfig.sh

Using /web/cognos11/jre/bin/java
06/06/2018,11:13:04,Dbg,Use Customized settings for font and color.
SLF4J: Class path contains multiple SLF4J bindings.
SLF4J: Found binding in [jar:file:/web/cognos11/bin/slf4j-nop-1.7.23.jar!/org/slf4j/impl/StaticLoggerBinder.class]
SLF4J: Found binding in [jar:file:/web/cognos11/configuration/utilities/config-util.jar!/org/slf4j/impl/StaticLoggerBinder.class]
SLF4J: See http://www.slf4j.org/codes.html#multiple_bindings for an explanation.
SLF4J: Actual binding is of type [org.slf4j.helpers.NOPLoggerFactory]
06/06/2018,11:13:10,Dbg,The original cogstartup.xml file is clear text. Don't back it up.

That is to say, it worked! I’ve often seen software packages install their own versions of Java. This is the first time I thought to take advantage of that. Wish I had thought of this approach during the Cognos 10 install!

Categories
Apache Linux Web Site Technologies

Turning Apache into a Redirect Factory

Intro
I’m getting a little more used to Apache. It’s a strange web server with all sorts of bolt-on pieces. The official documentation is horrible so you really need sites like this to explain how to actually do useful things. You needs real, working examples. In this example I’m going to show how to use the mod_rewrite engine of Apache to build a powerful and convenient web server whose sole purpose in life is for all types of redirects. I call it a redirect factory.

Which Redirects Will it Handle
The redirects will be read in from a file with an easy, editable format. So we never have to touch our running web server. We’ll build in support for the types of redirect requests that I have actually encountered. We don’t care what kind of crazy stuff Apache might permit. You’ll pull your hair out trying to understand it all. All redirects I have ever encountered fall into a relatively small handful of use cases. Ordered by most to least common:

  1. host -> new_url
  2. host/uri[Suffix] -> new_fixed_url (this can be a case-sensitive or case-insensitive match to the uri)
  3. host/uri[Suffix] -> new_prefix_uri[Suffix] (also either case-sensitive or not)

So some examples (not the best examples because I don’t manage drj.com or drj.net, but pretend I did):

  1. drj.com/WHATEVER -> http://drjohnstechtalk.com/
  2. www.drj.com -> http://drjohnstechtalk.com/
  3. drj.com/abcPATH/Preserve -> http://drjohnstechtalk.com/abcPATH/Preserve
  4. drj.com/defPATH/Preserve -> http://drjohnstechtalk.com/ghiPATH/Preserve
  5. drj.com/path/with/slash -> http://drjohnstechtalk.com/other/path
  6. drj.com/path/with/prefix -> http://drjohnstechtalk.com/other/path
  7. drj.net/pAtH/whatever -> https://drjohnstechtalk.com/straightpath
  8. drj.net/2pAtH/stuff?hi=there http://drjohnstechtalk.com/2straightpath/stuff?hi=there
  9. my.host -> http://regular-redirect.com/
  10. whatever-host.whatever-domain/whatever-URI -> http://whatever-new-host.whatever-new-domain/whatever-new-URI

All these different cases can be handled with one config file. I’ve named it redirs.txt. It looks like this:

# redirs file
# The default target has to be listed first
defaultTarget   D       http://www.drjohnstechtalk.com/blog/
# hosts with URI-matching grouped together
# available flags: "P" - preserve part after match
#                  "C" - exact case match of URI
 
# Begin host: drj.com:www.drj.com - ":"-separated list of applicable hostnames
/                       http://drjohnstechtalk.com/
/abc    P       http://drjohnstechtalk.com/abc
/def    P       http://drjohnstechtalk.com/ghi
/path/with/slash https://drjohnstechtalk.com/other/path
/path/with/prefix P  https://drjohnstechtalk.com/other/path
# end host drj.com:www.drj.com
 
# this syntax - host/URI - is also OK...
drj.net/ter             http://drjohnstechtalk.com/terminalredirect
drj.net/pAtH    C       http://drjohnstechtalk.com/straightpath
drj.net/2pAtH   CP      http://drjohnstechtalk.com/2straightpath
 
# hosts with only host-name matching
my.host                 http://regular-redirect.com/
www.drj.edu             http://education-redirect.edu/edu-path

The Apache configuration file piece is this:

# I really don't think this does anything other than chase away a scary warning in the error log...
RewriteLock ${APACHE_LOCK_DIR}/rewrite_lock
<VirtualHost *:90>
# Inspired by the dreadful documentation on http://httpd.apache.org/docs/2.0/mod/mod_rewrite.html
RewriteEngine on
RewriteMap  redirectMap prg:conf/vhosts/redirect.pl
#RewriteCond ${lowercase:%{HTTP_HOST}} ^(.+)$
RewriteCond ${redirectMap:%{HTTP_HOST}%{REQUEST_URI}} ^(.+)$
# %N are backreferences to RewriteCond matches, and $N are backreferences to RewriteRule matches
RewriteRule ^/.* %1 [R=301,L]
</VirtualHost>

Remember I split up apache configuration into smaller files. So that’s why you don’t see the lines about logging and what port to listen on, etc. And the APACHE_LOCK_DIR is an environment variable I set up elsewhere. This file is called redirect.conf and is in my conf/vhosts directory.

In my main httpd.conf file I extended the logging to prefix the lines in the access log with the host name (since this redirect server handles many host names this is the only way to get an idea of which hosts are popular):

...
    LogFormat "%{Host}i %h %l %u %t \"%r\" %>s %b \"%{Referer}i\" \"%{User-Agent}i\"" combined
...

So a typical log line looks something like the following:

drj.com 201.212.205.11 - - [10/Feb/2012:09:09:07 -0500] "GET /abc HTTP/1.1" 301 238 "http://www.google.com.br/url?sa=t&rct=j&q=drjsearch" "Mozilla/4.0 (compatible; MSIE 8.0; Windows NT 5.1; Trident/4.0; .NET CLR 2.0.50727; .NET CLR 3.0.4506.2152; .NET CLR 3.5.30729; .NET4.0C; .NET4.0E)"

I had to re-compile apache because originally my version did not have mod_rewrite compiled in. My description of compiling Apache with this module is here.

The directives themselves I figured out based on the lousy documentation at their official site: http://httpd.apache.org/docs/2.0/mod/mod_rewrite.html. The heavy lifting is done in the Perl script because there you have some freedom (yeah!) and are not constrained to understand all their silly flags. One trick that does not seem documented is that you can send the full URL to your mapping program. Note the %{HTTP_HOST}%{REQUEST_URI} after the “:”.

I tried to keep redirect.pl brief and simple. Considering the many different cases it isn’t too bad. It weighs in at 70 lines. Here it is:

#!/usr/bin/perl
# Copyright work under the Artistic License, http://www.opensource.org/licenses/Artistic-2.0
# input is $HTTP_HOST$REQUEST_URI
$redirs = "redirs.txt";
# here I only want the actual script name
$working_directory = $script_name = $0;
$script_name =~ s/.*\///g;
$working_directory =~ s/\/$script_name$//g;
$finalType = "";
$DEBUG = 0;
$|=1;
while (<STDIN>) {
  chomp;
  ($host,$uri) = /^([^\/]+)\/(.*)/;
  $host = lc $host;
# use generic redirect file
  open(REDIRS,"$working_directory/$redirs") || die "Cannot open redirs file $redirs!!\n";
  $lenmatchmax = -1;
  while(<REDIRS>) {
# look for alternate names section
    if (/#\s*Begin host\s*:\s*(\S+)/i) {
      @hostnames = split /:/,$1;
      $pathsection = 1;
    } elsif (/#\s*End host/i) {
      $pathsection = 0;
    }
    @hostnames = () unless $pathsection;
    next if /^#/ || /^\s*$/; # ignore comments and blank lines
    chomp;
    $type = "";
# take out trailing spaces after the target URL
    s/\s+$//;
    if (/^(\S+)\s+(\S{1,2})\s+(\S+)$/) {
      ($redirsURL,$type,$targetURL) = ($1,$2,$3);
    } else {
       ($redirsURL,$targetURL) = /^(\S+)\s+(\S+)$/;
    }
# set default target if specified. It has to come at beginning of file
    $finalURL = $targetURL if $type =~ /D/;
    $redirsHost = $redirsURI = $redirsURIesc = "";
    ($redirsHost,$redirsURI) = $redirsURL =~ /^([^\/]*)\/?(.*)/;
    $redirsURIesc = $redirsURI;
    $redirsURIesc =~ s/([\/\?\.])/\\$1/g;
    print "redirsHost,redirsURI,redirsURIesc,targetURL,type: $redirsHost,$redirsURI,$redirsURIesc,$targetURL,$type\n" if $DEBUG;
    push @hostnames,$redirsHost unless $pathsection;
    foreach $redirsHost (@hostnames) {
    if ($host eq $redirsHost) {
# assume case-insensitive match by default.  Use type of 'C' to demand exact case match
# also note this matches even if uri and redirsURI are both empty
      if ($uri =~ /^$redirsURIesc/ || ($type !~ /C/ && $uri =~ /^$redirsURIesc/i)) {
# find longest match
        $lenmatch = length($redirsURI);
        if ($lenmatch > $lenmatchmax) {
          $finalURL = $targetURL;
          $finalType = $type;
          $lenmatchmax = $lenmatch;
          if ($type =~ /P/) {
# prefix redirect
            if ($uri =~ /^$redirsURIesc(.+)/ || ($type !~ /C/ && $uri =~ /^$redirsURIesc(.+)/i)) {
              $finalURL .= $1;
             }
          }
        }
      }
    } # end condition over input host matching host from redirs file
    } # end loop over hostnames list
  } # end loop over lines in redirs file
  close(REDIRS);
# non-prefix re-direct. This is bizarre, but you have to end URI with "?" to kill off the query string, unless the target already contains a "?", in which case you must NOT add it! Gotta love Apache...
  $finalURL .= '?' unless $finalType =~ /P/ || $finalURL =~ /\?/;
  print "$finalURL\n";
} # end loop over STDIN

The nice thing here is that there are a couple of ways to test it, which gives you a sort of cross-check capability. Of course I made lots of mistakes in programming it, but I worked through all the cases until they were all right, using rapid testing.

For instance, let’s see what happens for www.drj.com. We run this test from the development server as follows:

> curl -i -H ‘Host: www.drj.com’ ‘localhost:90’

HTTP/1.1 301 Moved Permanently
Date: Thu, 09 Feb 2012 15:24:25 GMT
Server: Apache/2
Location: http://drjohnstechtalk.com/
Content-Length: 235
Content-Type: text/html; charset=iso-8859-1
 
<!DOCTYPE HTML PUBLIC "-//IETF//DTD HTML 2.0//EN">
<html><head>
<title>301 Moved Permanently</title>
</head><body>
<h1>Moved Permanently</h1>
<p>The document has moved <a href="http://drjohnstechtalk.com/">here</a>.</p></body></html>

And from the command line I test redirect.pl as follows:

> echo “www.drj.com/”|./redirect.pl

http://drjohnstechtalk.com/?

That terminal “?” is unfortunate, but apparently you need it to kill off any possible query_string.

You want some more? OK. How about matching a host and the initial path in a case-insensitive manner? No problem, we’re up to the challenge:

> curl -i -H ‘Host: DRJ.COM’ ‘localhost:90/PATH/WITH/SLASH/stuff?hi=there’

HTTP/1.1 301 Moved Permanently
Date: Thu, 09 Feb 2012 15:38:12 GMT
Server: Apache/2
Location: https://drjohnstechtalk.com/other/path
Content-Length: 246
Content-Type: text/html; charset=iso-8859-1
 
<!DOCTYPE HTML PUBLIC "-//IETF//DTD HTML 2.0//EN">
<html><head>
<title>301 Moved Permanently</title>
</head><body>
<h1>Moved Permanently</h1>
<p>The document has moved <a href="https://drjohnstechtalk.com/other/path">here</a>.</p></body></html>

Refer back to the redirs file and you see this is the desired behaviour.

We could go on with an example for each case, but we’ll conclude with one last one:

> curl -i -H ‘Host: DRJ.NET’ ‘localhost:90/2pAtHstuff?hi=there’

HTTP/1.1 301 Moved Permanently
Date: Thu, 09 Feb 2012 15:44:37 GMT
Server: Apache/2
Location: http://drjohnstechtalk.com/2straightpathstuff?hi=there
Content-Length: 262
Content-Type: text/html; charset=iso-8859-1
 
<!DOCTYPE HTML PUBLIC "-//IETF//DTD HTML 2.0//EN">
<html><head>
<title>301 Moved Permanently</title>
</head><body>
<h1>Moved Permanently</h1>
<p>The document has moved <a href="http://drjohnstechtalk.com/2straightpathstuff?hi=there">here</a>.</p>
</body></html>

A case-sensitive, preserve match. Change “pAtH” to “path” and there is no matching line in redirs.txt so you will get the default URL.

Creating exceptions

Eventually I wanted to have an exception – a URI which should be served with a 200 status rather than redirected. How to handle?

# Inspired by the dreadful documentation on http://httpd.apache.org/docs/2.0/mod/mod_rewrite.html
        RewriteEngine on
# just this one page should NOT be redirected
        Rewriterule ^/dontredirectThisPage.php - [L]
        RewriteMap  redirectMap prg:redirect.pl
        ... etc ...

The above apache configuration snippet shows that I had to put the page which shouldn’t be redirected at the top of the ruleset and set the target to “-“, which turns off redirection for that match, and make this the last executed Rewrite rule. I think this is better than a negated match (!) which always gets complicated.

Conclusion
A powerful redirect factory was constructed from Apache and Perl. We suffered quite a bit during development because of incomprehensible documentation. But hopefully we’ve saved someone else this travail.

References and related
This post describes how to massage Apache so that it always returns a maintenance page no matter what URI was originally requested.
I have since learned that another term used in the industry for rediect server is persistent URL (PURL). It’s explained in Wikipedia by this article: https://en.wikipedia.org/wiki/Persistent_uniform_resource_locator

Categories
Admin Internet Mail Linux SLES

Building sendmail on SLES

Intro
My sendmail binary built for SLES 10 SP 3 was not working well at all on SLES 11 SP1. It became apparent that libraries were not compatible so it was time to re-compile. I’ve documented that journey here. There were a few pitfalls along the way so I felt it was worth a blog post should anyone else ever need to do this.

February 2013 update
And now I’ve repeated the journey on SLES 11 SP2 – and ran into new problems! I’ll put that story in the appendix below.

Why Build?
Why build sendmail when you can find a package for it? For security it’s a good idea to run the latest version. It’s easier to defend during an audit. So when I look via zypper, I see it proposes me sendmail v 8.14.3:

# sudo zypper if sendmail
Refreshing service 'nu_novell_com'.
Loading repository data...
Reading installed packages...
 
Information for package sendmail:
 
Repository: SLES11-SP1-Pool
Name: sendmail
Version: 8.14.3-50.20.1
Arch: x86_64
Vendor: SUSE LINUX Products GmbH, Nuernberg, Germany
...

I go to sendmail.org and find that the latest version is actually 8.14.5. And that’s fairly typical. The distributed release is over a year old.

Where this can really matter is when a vulnerability comes out. If you can roll your own you can be the first on the block with that vulnerability fixed – not putting yourself at the mercy of a vendor busy with hundreds of other distractions. And I have seen this phenomenon in action.

Distributing Your Build
I’m mixing up the order here.

Once you have your sendmail built, what’s the minimum set of things you need to put it on other servers?

For one, you gotta have a database package with sendmail. For historical reasons I use sleepycat (I think formerly known as Berkeley db). Only it was gobbled up by Oracle. I don’t have a feel for what the future holds, though I fear decline. Sleepycat provides db. I grabbed the RPM from rpmfind.net:

db-4.7.25-1rt.x86_64.rpm.

This package has to go on each server where you will run sendmail if you are using db for your database. First issue in trying to install this RPM:

# sudo rpm -i db-4.7.25-1rt.x86_64.rpm
        file /usr/bin/db_archive from install of db-4.7.25-1rt.x86_64 conflicts with file from package db-utils-4.5.20-95.39.x86_64
        file /usr/bin/db_checkpoint from install of db-4.7.25-1rt.x86_64 conflicts with file from package db-utils-4.5.20-95.39.x86_64
        file /usr/bin/db_deadlock from install of db-4.7.25-1rt.x86_64 conflicts with file from package db-utils-4.5.20-95.39.x86_64
        file /usr/bin/db_dump from install of db-4.7.25-1rt.x86_64 conflicts with file from package db-utils-4.5.20-95.39.x86_64
        file /usr/bin/db_hotbackup from install of db-4.7.25-1rt.x86_64 conflicts with file from package db-utils-4.5.20-95.39.x86_64
        file /usr/bin/db_load from install of db-4.7.25-1rt.x86_64 conflicts with file from package db-utils-4.5.20-95.39.x86_64
        file /usr/bin/db_printlog from install of db-4.7.25-1rt.x86_64 conflicts with file from package db-utils-4.5.20-95.39.x86_64
        file /usr/bin/db_recover from install of db-4.7.25-1rt.x86_64 conflicts with file from package db-utils-4.5.20-95.39.x86_64
        file /usr/bin/db_stat from install of db-4.7.25-1rt.x86_64 conflicts with file from package db-utils-4.5.20-95.39.x86_64
        file /usr/bin/db_upgrade from install of db-4.7.25-1rt.x86_64 conflicts with file from package db-utils-4.5.20-95.39.x86_64
        file /usr/bin/db_verify from install of db-4.7.25-1rt.x86_64 conflicts with file from package db-utils-4.5.20-95.39.x86_64

I don’t really know if any of these conflicting files are used by the system. I also don’t know how to install “all but the conflicting files” in RPM. So we’ll try our luck and simply overwrite them:

# sudo rpm -i –force db-4.7.25-1rt.x86_64.rpm

Now no errors are reported.

We gotta make another kludge. sendmail, or at least the version I compiled, needs this db library in /usr/lib64, but the RPM puts it in /usr/lib. So…

# cd /usr/lib64; sudo ln -s ../lib/libdb-4.7.so

By the way, how did I decide that db has to be brought over? I fired up sendmail and got this error:

# sendmail
sendmail: error while loading shared libraries: libdb-4.7.so: cannot open shared object file: Error 40

Now with the sym link made I get a more pleasant application error:

# sendmail
can not chdir(/var/spool/clientmqueue/): Permission denied
Program mode requires special privileges, e.g., root or TrustedUser.

Continuing my out-of-order documentation(!), I create an init script in /etc/init.d and make sure sendmail is going to be started at boot:

# sudo chkconfig -s drjohnssendmail 35

I like to put the logs in /maillog:

# sudo rm /var/log/mail; sudo mkdir !$; sudo ln -s !$ /maillog

I like to have the logging customized a bit, so I modify syslog-ng.conf somewhat:

# DrJ attempt to define filter based on match of sm-mta
filter f_sm-mta     { match("sm-mta"); };
filter f_fs-mta     { match("fs-mta"); };

and

...
#destination mail { file("/var/log/mail"); };
#log { source(src); filter(f_mail); destination(mail); };
 
destination mailwarn { file("/var/log/mail/mail.warn" perm(0644)); };
log { source(src); filter(f_mailwarn); destination(mailwarn); };
 
destination mailerr  { file("/var/log/mail/mail.err" perm(0644) fsync(yes)); };
log { source(src); filter(f_mailerr);  destination(mailerr); };
 
#
# and also all in one file:
#
#
# and also all in one file:
#
destination mail { file("/var/log/mail/stat.log" perm(0644)); };
log { source(src); filter(f_sm-mta); destination(mail); };
log { source(src); filter(f_fs-mta); destination(mail); };

Followed by a

# sudo service syslog restart

Going Back to Compiling
So how did we compile sendmail in the first place, which was supposed to be the subject of this blog?

We downloaded the latest version from sendmail.org and unpacked it. Then we read the INSTALL file in the sendmail-8.14.5 directory for general guidance about the steps.

To make our ocmpilation configuration portable, we try to encapsulate our idiosyncracies in one file, devtools/Site/site.config.m4, which we create. Mine looks as follows:

dnl  DrJ config file for corporate mail delivery
dnl I am leaving out the ldap stuff because we stopped using it
dnl 
dnl which maps we will support - NEWDB is automatic if it finds the db libs
dnl in Linux NDBM is really GDBM, which isn't supported.  NEWDB support is not automatic
define(`confMAPDEF',`-DNEWDB -DMAP_REGEX')
dnl Berkeley DB is here...
dnl this doesn't work, exactly - put the db libs directly into /usr/lib
APPENDDEF(`confLIBDIRS',`-L/usr/local/ssl/lib -L/usr/lib64')
dnl libdb-4 looks like the sleepycat library on Linux
APPENDDEF(`confLIBS',`-ldb-4')
APPENDDEF(`confINCDIRS',`-I/opt/local/include -I/usr/local/ssl/include')
dnl where to put smrsh and mail.local programs
define(`confEBINDIR', `/opt/mail/bin')
dnl where to install include files
define(`confINCLUDEDIR', `/opt/mail/include')
dnl where to install library files
define(`confLIBDIR', `/opt/mail/lib')
dnl man pages
define(`confMANROOT', `/opt/local/man/cat')
dnl unformatted man pages
define(`confMANROOTMAN', `/opt/local/man/man')
dnl the sendmail binary goes into MBINDIR
define(`confMBINDIR', `/opt/mail/bin')
dnl programs only executed by root go to sbin
define(`confSBINDIR', `/opt/mail/sbin')
dnl shared library directory
define(`confSHAREDLIBDIR', `/opt/mail/lib')
dnl user-executable pgms, newaliases, mailq, hoststat, etc
define(`confUBINDIR', `/opt/mail/bin')
dnl TLS support 
APPENDDEF(`conf_sendmail_ENVDEF', `-DSTARTTLS')
APPENDDEF(`conf_sendmail_LIBS', `-lssl -lcrypto')

I’ll explain a bit of this file since of course it is critical to what we are trying to do. I arrived at its current state from a little experimentation, so I don’t know the full explanation of some of the settings. What it’s saying is that we use the NEWDB, which uses that db we spoke of earlier for our maps. I like to install the binaries into /opt/mail/bin. We like to have the option to run TLS.

With that set we run the compile:

# cd sendmail; sh ./Build

and it spits out some errors to the screen which indicate we’re missing some SSL headers. We get them with:

# sudo zypper source-install openssl

Now with any luck it will fully compile.

At this point it helps to create some of the target directories:

# sudo mkdir -p /opt/mail/{bin,sbin} /opt/local/man/cat{1,5,8}

And we create a user account, smmsp, with uid and gid of 225, and a group with the same name.

And then we can install it:

# sudo sh ./Build install

The install should work, mostly. But makemap doesn’t get put in /opt/mail for some reason. So you have to copy it by hand from sendmail-8.14.5/obj.Linux.2.6.32.12-0.7-default.x86_64/makemap to /opt/mail/sbin, for instance. You really need to have a makemap.

Finally, I suggest to recursively copy the cf directory:

# cp -r sendmail-8.14.5/cf /opt/mail

This way you have a pretty relocatable set of files under /opt/mail.

Appendix: Building on SLES 11 SP2
I thought this would be a breeze. Just look at my own blog posting above! Not so fast – that approach didn’t work at all.

You have to appreciate that under SLES 11 SP1 we needed a few key packages that aren’t very common:

– libopenssl-devel-0.9.8h-30.27.11
– zlib-devel-1.2.3-106.34

We pulled them from the SDK DVD. Well, turns out there is no SDK DVD for SLES 11 SP2! Novell, probably in one of those beloved cost-saving measures, no longer releases an SDK DVD. What to do?

Well, I found what not to do. I began copying key files from my SLES 11 SP1 installation, like libcrypto.a, libssl.a, /usr/include/ssl. This all helped to reduce the number of errors. But at the end of the day there was an error I couldn’t chase away no matter what:

/usr/src/packages/BUILD/openssl-0.9.8h/crypto/comp/c_zlib.c:235: undefined reference to `inflate'

Meanwhile the resourceful sysadmin found those development packages. He told me about SUSE Studio, which allows you to build your own distribution. He looked for those development packages in the distribution, found and installed them:

$ rpm -qa|grep devel

libopenssl-devel-0.9.8j-0.26.1
zlib-devel-1.2.3-106.34
zlib-devel-32bit-1.2.3-106.34

Then my build went through fine. Whew!

Conclusion
I could go on and on about details in the setup, but the scope here is the compilation, and we’ve covered that pretty well.

Once you get over the pain of compilation setup, sendmail runs great and is a great MTA.

Categories
Admin IT Operational Excellence Linux Proxy Web Site Technologies

The IT Detective Agency: intermittent web page not found error

Intro
One of the high arts of IT is system integration, and an important off-shoot of this is acquisitions. We are involved in integrating a new location, which, unfortunately, we do not yet have full access to. The local networking is still provided by their vendor, not ours and this makes troubleshooting all the more difficult.

The Details
So the word begins to spread that users at this site are having intermittent problems accessing some of our secure web sites. As it was described to me, they can load the page in their browser for, say five straight times, get a simple Internet Explorer cannot display the web page error, and the sixth time (or whenever) it will load properly. All other connectivity was working. No one else at other locations was having this problem with this web site. More than strange, right?

In drjohn’s perfect IT world, problem reproducibility is critical to resolution, but we simply didn’t have it this time. I also could not produce the problem myself, which means relying on other people.

I’m not sure if we tried to contact their vendor or not at first. But if we had I’m sure they would have denied having anything to do with it.

So we got one of our confederates, Tim, over to this location and we hooked him up with Wireshark so he could get take a packet trace when the failure occurs. It wasn’t long before Tim reproduced the error and emailed us the packet capture.

In the following the PC has IP address 10.200.23.34, the web server is at 10.4.5.6. The Linux command used to look at the capture file is:

# tcpdump -A -r bodega-error.cap port 443 > /tmp/dump

1 15:54:27.495952 IP 10.200.23.34 > 10.4.5.6.https: S 2803722614:2803722614(0) win 64240 <mss 1460,nop,wscale 0,nop,nop,sackOK>
2 15:54:27.496309 IP 10.4.5.6.https > 10.200.23.34: S 3201081612:3201081612(0) ack 2803722615 win 5840 <mss 1432,nop,nop,sackOK>
3 15:54:27.496343 IP 10.200.23.34 > 10.4.5.6.https: . ack 1 win 64240
4 15:54:27.497270 IP 10.200.23.34 > 10.4.5.6.https: P 1:82(81) ack 1 win 64240
5 15:54:27.497552 IP 10.4.5.6.https > 10.200.23.34: . ack 82 win 5840
6 15:54:30.743827 IP 10.4.5.6.https > 10.200.23.34: P 1:286(285) ack 82 win 5840
..S.......^M..i.P.......HTTP/1.0 200 OK^M
Cache-Control: no-store^M
Pragma: no-cache^M
Cache-Control: no-cache^M
X-Bypass-Cache: Application and Content Networking System Software 5.5.17^M
Connection: Close^M
^M
<HTML><HEAD><META HTTP-EQUIV="REFRESH" CONTENT="0;URL=https://10.4.5.6/"></HEAD><BODY>
</BODY></HTML>
 
7 15:54:30.744036 IP 10.200.23.34 > 10.4.5.6.https: F 82:82(0) ack 286 win 63955
8 15:54:30.744052 IP 10.4.5.6.https > 10.200.23.34: F 286:286(0) ack 82 win 5840
9 15:54:30.744077 IP 10.200.23.34 > 10.4.5.6.https: . ack 287 win 63955
10 15:54:30.744289 IP 10.4.5.6.https > 10.200.23.34: . ack 83 win 5840

The output was scrubbed a bit of meaningless junk characters and I added serial packets numbers in the beginning by hand because I don’t (yet) know how to do that with tcpdump!

What, It’s Encrypted – what can you even learn from a trace?
Yeah, an SSL stream sure adds to the already steep challenges we faced in this problem. There just isn’t much to work with. But it is something. I’m about to say what I noticed in this packet trace, but for it to be meaningful you need to know like I did that the web server is situated almost four thousand miles from the user’s location.

The first packet is a SYN from the PC to web server on TCP port 443. So far so good. In fact packets one – three constitute the three-way handshake in TCP.

Although SSL is encrypted, the beginning of the protocol communication should show the SSL cipher being chosen. Unfortunately, tcpdump doesn’t seem to have the smarts to show any of this. So I got myself ssldump. On Ubuntu:

# sudo apt-get install ssldump

did the trick. Then run this same capture file through ssldump, which has very similar arguments to tcpdump:

# ssldump -r bodega-error.cap port 443

New TCP connection #1: 10.200.23.34(2027) <-> 10.4.5.6(443)
1 1  0.0013 (0.0013)  C>S SSLv2 compatible client hello
  Version 3.1
  cipher suites
  TLS_RSA_WITH_RC4_128_MD5
  TLS_RSA_WITH_RC4_128_SHA
  TLS_RSA_WITH_3DES_EDE_CBC_SHA
  SSL2_CK_RC4
  SSL2_CK_3DES
  SSL2_CK_RC2
  TLS_RSA_WITH_DES_CBC_SHA
  SSL2_CK_DES
  TLS_RSA_EXPORT1024_WITH_RC4_56_SHA
  TLS_RSA_EXPORT1024_WITH_DES_CBC_SHA
  TLS_RSA_EXPORT_WITH_RC4_40_MD5
  TLS_RSA_EXPORT_WITH_RC2_CBC_40_MD5
  SSL2_CK_RC4_EXPORT40
  SSL2_CK_RC2_EXPORT40
  TLS_DHE_DSS_WITH_3DES_EDE_CBC_SHA
  TLS_DHE_DSS_WITH_DES_CBC_SHA
  TLS_DHE_DSS_EXPORT1024_WITH_DES_CBC_SHA
  Unknown value 0xff
Unknown SSL content type 72
1    3.2480 (3.2467)  C>S  TCP FIN
1 2  3.2481 (0.0000)  S>CShort record
1    3.2481 (0.0000)  S>C  TCP FIN

The way to interpret this is that 0.0013 s into the TCP port 443 communication the cipher suites listed above were sent by the PC to the server. This corresponds to our packet number 4 in the trace file.

Using Wireshark to look at the trace is a lot more convenient – it provides packet numbers, timing, decodes packets and displays the SSL ciphers. But I wanted to show that it _could_ be done with text-based tools.

Look at the timings more closely. In the tcpdump output, packet 2, the SYN ACK, comes 1 ms after the SYN. But given the distances involved between PC and server, the SYN ACK should have come more like 100 ms later, at least. Similarly packet 5, which is an ACK, comes less than 1 ms after packet 4. A 1 ms ACK? Physically impossible.

I have seen this behaviour before – on our own load balance – which I know employs some TCP optimization tricks. So I concluded that they must have physically present at this site some kind of appliance which is doing TCP optimization. It can only provide blank ACKs in its rapid-fire responses since it can’t know what data the server is really going to respond with. That might all be OK. But I’m pretty sure the problem lies between packets 5 and 6. 5 is one of those meaningless rapid-fire empty ACKs generated by the local router. But the PC has just sent a wish list of SSL ciphers in packet 4. It needs to be responded to by the server which has to finish setting up the SSL session.

But that critical packet from the server never arrives. Perhaps even some of the SSL handshake is secretly completed between the local router and the server. Who knows? I have heard of man-in-the-middle devices that decrypt SSL sessions. And packet 6 contains fairly inappropriate content. It almost does look like it has been manufactured by a man-in-the-middle device. Its telling the browser to do a redirect to the same site, except specified by IP address rather than FQDN. And that doesn’t make a lot of sense. The browser likely realizes that this amounts to a looping redirect request so at that point it probably decides to cut its losses and FIN the connection in packet 7.

I traced my own PC hitting this same web server. Now I know we don’t have any of these optimizing devices between me and the web server. I don’t have time to show the results here, but to summarize, it looks rather completely different from the trace above. The ACK packets come back in about 100 ms or so. There is no delay of three seconds. The cipher proposals are responded to in a timely fashion. There is no redirect.

Their Side of the Story
We did get to hear back from the vendor who supports the LAN/WAN. They said they were running WCCP and diverting traffic to a proxy server. This was the correct behaviour before we hooked our infrastructure to this site, but is no longer. They realized this was probably a bad thing and took corrective action to turn off WCCP for destinations in the internal network 10.0.0.0/8.

Conclusion
Shutting off WCCP, which diverted web site requests to an old proxy server, fixed the problem.

Case closed.

Unsolved Mysteries
I wish we could tie all the loose ends neatly up, but there are too many players involved. We’ll never really know why the problem was intermittent, for instance. Or why some secure web sites could be accessed without any issue whatsoever throughout this ordeal.

WCCP, Web cache Communication Protocol, is a Cisco-developed routing protocol to transparently intercept traffic destined for web servers. More information can be found on it in wikipedia.

It bothers me that after the SSL session was initiated the dump showed the source, unencrypted, of the HTML redirect packet. Why wasn’t that encrypted? Perhaps the WCCP-invoked proxy server was desperately trying to help the PC recover from an unrecoverable situation and manufactured that HTTP-EQUIV REFRESH… to try to force the PC to choose a web site that might work. The fact that it was sent unencrypted over a channel that should have been encrypted was probably even the death bell that triggered the browser to think this makes no sense at all and is even a violation of security, I’m getting out of here.

Categories
Admin Internet Mail Linux

The IT Detective Agency: generated email goofs attachments

Intro
Today was a busy day! A rather expert user asks for my advice about emails being generated from his own Unix system, by his own processes, that occasionally come out with the encoding of the attachments showing up in the body of the message.

I am not a message formatting expert, or at least I wasn’t prior to this question today. Bbut if a sendmail expert can’t provide an answer, who will? So anyways, this user, let’s call him Rob forwards me this email:

Dear DrJohn,
 
I was wondering if you have some idea as to why sometimes the attachment 
is getting included in the text of the email instead of being recognized as attachment, 
see example below.
 
Any pointers would be helpful, as this makes it at the least cumbersome 
to open the attachment by cutting , pasting the text attachment part 
as file and uudecoding it into binary before it can be opened for content view.
 
Regards,
Rob
 
Mime-Version: 1.0
Content-Type: multipart/mixed;
                 boundary="----=_Part_168946_477699193.1322837415283"
X-Mailer: sendmsg
 
------=_Part_168946_477699193.1322837415283
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit
 
Dear Team;
 
While executing the flow, In.process:RapidProductMovementReport_New, the following exception occurred.
 
java.lang.Exception: java.sql.SQLException:ORA-12899: value too large for column "B2B_RAPID"."P_MOVEMENT_DETAIL"."UNIT_OF_MEASURE_TYP" (actual: 3, maximum: 2)
 
 
Error Dump :
com.wm.lang.flow.FlowException:java.lang.Exception: java.sql.SQLException:ORA-12899: value too large for column ... (actual: 3, maximum: 2)
 
 
Pipeline values (see attachment)
 
Caller: Rapid_In.process:RapidRouter
 
Stack: %serviceStack%
------=_Part_168946_477699193.1322837415283
Content-Type: application/gzip; name=pipeline.xml.gz
Content-Transfer-Encoding: base64
Content-Disposition: attachment; filename=pipeline.xml.gz
 
H4sIAAAAAAAAAOy9bXPiSJr3+/rMp8hwbEzMHbFFKVOpp57umpVBBp0SkkcS5WHeVLhtutr32sbH
dtV0f/uTEmDLIMBISErBf2O3to2SJJUP18PvujLz53/8cXdLfkwen26m97+c0I5yQib3V9Prm/tv
v5yM4rMP5sk/Pv3l5y+Xt98nT68FmSj46S+E/PwjeUDuL+8mv5xcT6++303un+M/HyYnn84fp9ff
r56H0x+T5MNw8jB9fP75Y/qFla/eX/72fOlNvzl3lze39vX14+Tp6eTTKTu1r+9u7v/n18un3zpX
07t1X7++eZxcPYuWnXxy/dNg5PfWlbydfvvt5naS/HHy6eP04fnjfya/3k2ef59eP328fHi4vbm6
TCoS5Z4.......

Hmm. So what do you think? I have often seen these Mime-type headers and paid absolutely no attention to them. When things work what does it matter? But now they’re not working and it does matter.

First I tried to reproduce the problem using a technique I had gleaned looking over the shoulders of some Unix admins. I knew they had an easy method to send attachments from the command line of their Unix systems so I walked over and asked them how they did it. Sure enough, it was dead easy. It goes something like this:

# uuencode file.txt file.txt|mailx -s “here is your attachment” recipient_address

I rolled up my sleeves, set recipient_address to a valid SMTP mail address. Sure enough. I got it as an attachment in Lotus Notes. But it’s an ugly attachment and doesn’t have any of the nice MIME formatting about it. so it’s probably a bit of luck that my MUA (mail user agent) understands that I mean to create an attachment. I don’t think all MUAs will do that, unless it’s following a more obscure RFC which I’m not aware of.

The original source of the message looks like this (my attachment is called cogstartup.xml.gz):

Date: Fri, 2 Dec 2011 14:54:34 -0500
From: ...
Message-Id: ...
To: ...
Subject: test using uuencode
X-MIMETrack: Itemize by SMTP Server on ... (Release 7.0.4|March 23, 2009) at
 02/12/2011 20:54:36,
		 Serialize by Notes Client ... (Release 8.5.1|September
 28, 2009) at 12/02/2011 04:12:40 PM,
		 Serialize complete at 12/02/2011 04:12:40 PM
X-TNEFEvaluated: 1
 
begin 644 cogstartup.xml.gz
M'XL(""57L4X``V-O9W-T87)T=7`N>&UL`.U]VW;;R+'H^WP%HA?-G$5)MN<2
MCW?&>].2/:/$NL24,\EY\0(!D$0,`@P`2N9\_:E+W]$@`(J2LG-&*QE+)-!=
M755=MZZ...

Very different from what we saw above, right? None of those MIME-related headers seem to be present. So I decided that uuencode is too primitive to reproduce the problem.

I was next going to try to generate an email with the help of a PERL module, like MIME::Lite. But I decided that was too much trouble as I would need to download it and install it first!

So I ventured to see if I could get lucky and figure out the problem by educating myself on the standard, without wasting too much time. The relevant RFC seems to be 1341. I prefer the older RFCs because I suppose they’re shorter – easier to understand because life was simpler in those days! Once yuo parse through the verbiage and repitition, there wasn’t much to it. In particular, it mentioned that the Header

Mime-Version: 1.0

has to be included amongst the header fields. If it is not, the correct behaviour for a MUA is to interpret all the encodings and stuff as just part of a regular body text, which it will display to the user.

I ran a test using sendmail as my sending MUA from a Linux server. With the sendmail agent you can add headers, at least that’s how I remember it:

# sudo sendmail -v recipient< tst where tst is a file I created that starts with the line Mime-Version: 1.0 Yes, indeedy. I received it and my MUA interpreted the attachment as an attachment, displaying the attachment name and the appropriate icon type for it! Now put just a blank line at the top of my tst file, pushing all the rest of the stuff down by a line, and the behaviour is completely different. Then my MUA treats everything as literal body text, just as the old RFC says it must, and it looks just the way it did when Rob forwarded it to me. Conclusion
I explained to Rob that he must sometimes be introducing an extra line above the Mime-Verison header, which would cause this problem.

He thanked me.

Case closed!

Categories
Admin IT Operational Excellence Linux SLES

The IT Detective Agency: Cognos stopped working

Intro
Here’s another in our continuing exciting IT drama. A user reports that her Cognos app stopped working. She’s in charge of the Cognos application servers, I run the Cognos gateway on a Linux server. I have almost no working knowledge of Cognos. I learned just enough to get the gateway installed and configured on Linux, specifically SLES. Cognos is used for business intelligence reports and is now owned by IBM.

The Details
The home page came up just fine, so I knew the web server – Apache, of course – was working. I know I hadn’t changed anything on the gateway. She also says that she hadn’t changed anything on the dispatcher. So she asks me to save the config. It’s an X application. I run cogconfig.sh, which by the way is in COGNOS-INSTALL_DIR/bin64, not COGNOS-INSTALL_DIR/bin, contrary to the documentation for Linux. I cannot save the config. She asks me to export it. I can’t do that either! I get the error

CAM-CRP-1057 unable to generate the machine specific symmetric key.

She asks me to delete the keypairs. These are in the directories COGNOS-INSTALL_DIR/configuration/{signkeypair,encryptkeypair}. So I clear out those. Still I cannot save or export the configuration. I quickly switch to a Solaris server which we had hoped to retire in order to get a working gateway while we mulled the problem over.

Over the next days I checked to see if Java had changed. Getting a working JRE was a little tricky on SLES. Nothing had changed. After the system admin came back from vacation the next week I asked if by chance. The last log showed he was logged in at the time. He admits to changing one thing.

He changed the system name. This system has multiple interfaces and a unique hostname for each interface. The hosts file in /etc/hosts included entries for each of the interface IPs. Seeing there were no other changes I concluded that this little innocent act was enough to kill the communication. Note that he did not change any of the routing, however. When you’re dealing with encryption, it can be that the system name is significant. So when those keys were initially generated they were tied to that name and would only work with that original hostname. At least that is my reverse engineering of the matter. Cognos is a pretty closed system so it’s hard to pin down more precisely what is going on.

Conclusion
The hostname was changed back to the original name. Sure enough, now I can export the config and most importantly, save it without any errors.

Case closed!

Lessons Learned
Well, avoiding finger-pointing and quick judgements was helpful in this case. Of course I suspected she actually had done something to the dispatcher, but I behaved as though the problem might be on my side. We treated each other professionally while the system was down and we had no clue why. That was very helpful.

Categories
Admin IT Operational Excellence Linux SLES

The IT Detective Agency: the case of the messages from mars

Intro
Today we got a “funny” message on our SLES 11 server in the /var/log/warn file. You might think that Martians have landed!

The Details
Specifically this:

Nov 9 10:54:19 drjohn24 kernel: [72397.088297] martian source 10.120.2.24 from 10.0.0.3, on dev eth1
Nov 9 10:54:19 drjohn24 kernel: [72397.088300] ll header: 78:e7:d1:7b:25:32:00:a0:8e:a8:8e:b3:08:00

Every time I pinged 10.120.2.24 (drjohn24) from 10.0.0.3 it would produce those two lines in the warn and messages file. More worrisome, I could not ssh from one host to the other. I could ssh from a host on the local network to drjohn24. We observed this behaviour even with the firewall disabled. Strange, right?

One more thing to note: drjohn24 has two network interfaces and various routes defined.

The Solution
It didn’t take too long to get to the bottom of this. We set up the routes wrong. We meant to create a default route out of eth0, which was right, and a net-10 route for eth1, which we specified incorrectly. Do

netstat -rn

to show all routes. I had this:

Kernel IP routing table
Destination     Gateway         Genmask         Flags   MSS Window  irtt Iface
10.120.2.0      0.0.0.0         255.255.255.128 U         0 0          0 eth1
10.120.3.0      0.0.0.0         255.255.255.0   U         0 0          0 eth0
169.254.0.0     0.0.0.0         255.255.0.0     U         0 0          0 eth0
10.0.0.0        10.120.2.1      255.255.255.128 UG        0 0          0 eth1
128.0.0.0       0.0.0.0         255.0.0.0       U         0 0          0 lo
0.0.0.0         10.120.3.1      0.0.0.0         UG        0 0          0 eth0

Do you see the error? We put the mask on the 10.0.0.0 the same as we put on the interface and that’s not what we wanted.

The corrected version looks like this:

Destination     Gateway         Genmask         Flags   MSS Window  irtt Iface
...
10.0.0.0        10.120.2.1      255.0.0.0       UG        0 0          0 eth1
...

Conclusion
So what was happening is that the inbound packet from 10.0.0.3 was arriving at eth0 as we intended. But SLES 11 is now clever enough to realize, based on its routing table, that that is not the expected interface where a packet with that source IP should arrive. It should have arrived at eth1 because of the default route. No other static route was more specific for 10.0.0.3 due to our error. And apparently even with firewall turned off, SLES gets very defensive at this point. I’m not sure if it was sending return packets out of eth1 or not, because I kept looking for them out of eth0!

Once we corrected the routes the inbound packet arrived at eth0 and was returned with an answer packet from eth0 and the martian messages went away.

The martian message thing is a little obscure, and at the time more a distraction than anything else as we had to research what that meant. I guess for the future we’ll instantly know. It’s very similar to defining network topology on your firewalls in an anti-spoofing defense.

Case closed!