I’m running SLES 11.0 on a dual quad core i7 machine. I’ve been running into several problems lately, not the least of which my gui displays have stopped working. However, this is a new problem. I recently tried to install something, and found that there was no more disk space available, even though there should have been several hundred gigs left. I did a search for large files and there were two, from the last couple days:
-rw-r----- 1 root root 447G Oct 12 09:00 acpid-20101012
-rw-r----- 1 root root 63G Oct 12 16:03 acpid-20101013
500 gigs??
When I checked the processes acpid and syslog-ng are both running like crazy on a couple of cores.
The contents of the files appear to be this message over and over again:
Oct 12 16:03:36 cgs-cytosine acpid: can’t accept client: Too many open files
Oct 12 16:03:36 cgs-cytosine acpid: can’t accept client: Too many open files
Oct 12 16:03:36 cgs-cytosine acpid: can’t accept client: Too many open files
Oct 12 16:03:36 cgs-cytosine acpid: can’t accept client: Too many open files
Oct 12 16:03:36 cgs-cytosine acpid: can’t accept client: Too many open files
Oct 12 16:03:36 cgs-cytosine acpid: can’t accept client: Too many open files
Oct 12 16:03:36 cgs-cytosine acpid: can’t accept client: Too many open files
Oct 12 16:03:36 cgs-cytosine acpid: can’t accept client: Too many open files
Oct 12 16:03:36 cgs-cytosine acpid: can’t accept client: Too many open files
500 gigs of that message! On a different note, the machine compressed that entire file down to 20 megs. Sweet compression.
Any idea what is going on, or how I could at least stop it from using up so many resources?
Ah, cgs-cytosine is just the name of the server. This problem only started recently, while I was trying to fix another problem. Updated kernel etc. Unfortunately I don’t have a gui, so I’m restricted to commandline access (currently ssh-ing)
And yeah, I deep-sixed the 500gbs of that informative message
Hmm, well it seems to just hang. I’ve left it running for about 5 min, but it hasn’t produced any output. It might be worth noting that the /etc/syslog-ng and /sbin/acpid processes are still constantly running between 60%-100% and 25%-50% respectively, while the hald-addon-acpi: listening acpid socket [truncated here] process isn’t using any resources (waiting for the other processes to complete??)
On 2010-10-15 19:36, joshearl1 wrote:
>
> Hmm, well it seems to just hang. I’ve left it running for about 5 min,
> but it hasn’t produced any output. It might be worth noting that the
> /etc/syslog-ng and /sbin/acpid processes are still constantly running
There can not be a process named “/etc/syslog-ng”, that’s a configuration file.
> between 60%-100% and 25%-50% respectively, while the hald-addon-acpi:
> listening acpid socket [truncated here] process isn’t using any
> resources (waiting for the other processes to complete??)
They are probably busy writing all those messages per second. As a hack, you could filter them out
so that they don’t get written - later try to find out the root cause.
You should ask in the SLES forum, this might be something specific to that distribution.
–
Cheers / Saludos,
Carlos E. R.
(from 11.2 x86_64 “Emerald” at Telcontar)
Recently my server got the same problem and I’m stuck now, I have no idea about what causes that problem. If I restart the acpid process it seems to solve the problem however after aproximately few hours the problem returns.
Any news about that problem from 2010 (which is the creation year of this thread ;))
