asterisk users - Jan 2004 - FW: Memory problem

Dear all,

I have had an experience which I would run by all of you to see if this is
normal.

I am running a few asterisk servers with 512M RAM memory, and as I have
mentioned in previous notes, I have experienced frequent crashes when faced
with more than 15-20 simultaneous calls. I have tried to find out if it
could be due to (a) Xeon chip running HT, (b) old Kernel version 2.4.18-3,
(c) old redhat linux version 7.3, (d) H323 library pwlib and openh323
versions which are 1.5.2 and 1.12.2 respectively among many other
parameters. So far, unfortunately, the matter has not been resolved.
However, I have noticed that the memory usage on each server has built up
with time after the server being rebooted. I have complained about using
close to 500M even when there were very few calls on the server but nobody
seemed to be able to let me know if they were running at high memory usages
except for Jesse who was telling me that his memory usages have always been
low. Very recently, I noticed that after I rebooted the servers, the memory
usage would start at about 80 M and even after started the Asterisk threads,
I was running at about 100 M and even when there were calls, I was running
at about 100M-150M, but then after hours it would start to build up to 200M
and then 250M and then....finally close to 500M even after I stopped the
Asterisk threads, almost like there is a memory leak somewhere.

I wonder if that is normal, if someone can please tell me, or if not normal,
what could be the cause to it and how should this be rectified.

Thanks alot


Tom
---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.563 / Virus Database: 355 - Release Date: 1/17/2004

---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.563 / Virus Database: 355 - Release Date: 1/17/2004

asterisk-users-admin@lists.digium.com <> wrote:
> I am running a few asterisk servers with 512M RAM memory, and
> as I have
> mentioned in previous notes, I have experienced frequent
> crashes when faced
> with more than 15-20 simultaneous calls. I have tried to find
> out if it
> could be due to (a) Xeon chip running HT, (b) old Kernel
> version 2.4.18-3,
> (c) old redhat linux version 7.3, (d) H323 library pwlib and openh323
> versions which are 1.5.2 and 1.12.2 respectively among many other
> parameters. So far, unfortunately, the matter has not been resolved.
> However, I have noticed that the memory usage on each server
> has built up
> with time after the server being rebooted. I have complained
> about using
> close to 500M even when there were very few calls on the
> server but nobody
This is a linux question not an asterisk question. I have seen very
recent threads on here which explained what was happening and why, you
should try to review them.

I think you will find that if you reboot and don't start asterisk,
over-night your memory usage will increase to around 500M.
Most likely, during the night is when your cron scripts run, some of
these traverse the entire filesystem (update/locate) and prompt the OS
to save some of the FS contents in memory (ie cache).

In any case, it is normal to use almost all of your memory. If you
really have doubts, then look at ps aux which will show you memory usage
per process.

And, this is unlikely a asterisk problem and more an OS
mis-understanding.

Regards,
Adam

 --
Adam Goryachev
Website Managers
Ph:  +61 2 9345 4395                        adam@websitemanagers.com.au
Fax: +61 2 9345 4396                        www.websitemanagers.com.au