SQL Anywhere for Linux - Unexpected db engine shutdown

Author

Unexpected db engine shutdown

Jeffrey Fenningsdorf

2005-04-05, 8:03 pm

I'm having a odd problem that's been going on for a while now. Were
currently running Version 7.0.4.3517 on a Gentoo server with the
2.4.28-hardened-r2 hardened sources. All has been fine for a while
since we switched over from RedHat to Gentoo. But recently my problem
has cropped up again. The db engine will unexpectedly shutdown with no
errors in the console log, no core dump - nothing. If I turn on the
debugging it kills the app and makes the server useless. I've been
running 7.0.4 since RedHat 7.1 and I wrote a little script to monitor
the db engine process and restart it when this would happen, because you
know it's the server or the network :). My questions are....

1. Could this be a bug in 7.04 and should we upgrade to 9?
2. Is this a programming error?
3. Could this be the server setup?

I'm very doubtfully of #3 because I've run this db engine in many
different environments with many different types of hardware. As we've
grown so has my selection of hardware. I'm currently running a Compaq
with Dual 1.13 PIII's, and two Dells with Dual Xeon 3.06GHz. I'm
getting no where with this so any suggestions would be appreciated.

Jeffrey Fenningsdorf
Network Engineer
Professional Receivables Control

Mark Culp

2005-04-05, 8:03 pm

First let me say that I am astonished that the 7.0.4 server even starts on
2.4.28 since 7.0.4 was never tested on this kernel (or even anything even
close to it), and given the nature of how the Linux kernels have not been
very good about being upward-compatible (at least in our experience).
Either the later kernels are better at this, or we're getting better at
anticipating kernel changes? :-)

Without knowing more about why the engine is disappearing, there is no
way of even taking a guess as to why it is disappearing.

Some things to try:
- make sure that ulimit -c is set to unlimited for the server process
- make sure that the server has write permissions to the root / directory
(since cores will be written to the cwd of the process and a daemon
server does a "chdir /" when it starts).
This might give you a core, and if it does, there might be some information
there that could help determine the cause of the disappearance.
--
Mark Culp
ASA Research and Development
-------------------------------------------------------------------------
** Whitepapers, TechDocs, bug fixes are all available through the **
** iAnywhere Developer Community at http://www.ianywhere.com/developer **
-------------------------------------------------------------------------

Jeffrey Fenningsdorf wrote:
>
> I'm having a odd problem that's been going on for a while now. Were
> currently running Version 7.0.4.3517 on a Gentoo server with the
> 2.4.28-hardened-r2 hardened sources. All has been fine for a while
> since we switched over from RedHat to Gentoo. But recently my problem
> has cropped up again. The db engine will unexpectedly shutdown with no
> errors in the console log, no core dump - nothing. If I turn on the
> debugging it kills the app and makes the server useless. I've been
> running 7.0.4 since RedHat 7.1 and I wrote a little script to monitor
> the db engine process and restart it when this would happen, because you
> know it's the server or the network :). My questions are....
>
> 1. Could this be a bug in 7.04 and should we upgrade to 9?
> 2. Is this a programming error?
> 3. Could this be the server setup?
>
> I'm very doubtfully of #3 because I've run this db engine in many
> different environments with many different types of hardware. As we've
> grown so has my selection of hardware. I'm currently running a Compaq
> with Dual 1.13 PIII's, and two Dells with Dual Xeon 3.06GHz. I'm
> getting no where with this so any suggestions would be appreciated.
>
> Jeffrey Fenningsdorf
> Network Engineer
> Professional Receivables Control

Sponsored Links