Hi,
today, by lucky chance I realized that nmbd is eating all my memory.
This is a production server running Linux 2.0.33 and Samba 1.9.18p1.
After running for 9 days, nmbd seems to be using 23 MBytes of memory (!!!)
as shown by top (sorted by memory usage):
8:21am up 34 days, 7:57h, 1 user, load average: 0.06, 0.04, 0.00
56 processes: 55 sleeping, 1 running, 0 zombie, 0 stopped
CPU states: 0.9% user, 3.7% system, 0.0% nice, 95.3% idle
USER PID %CPU %MEM NI VSZ RSS SHRD TT STAT TIME COMMAND
root 9806 0.0 33.4 0 23564 21152 480 ? S 0:19 /usr/sbin/nmbd -
proxy 30619 0.0 17.2 0 14408 10912 548 ? S 1:07 squid -D -s -f /
[snip]
Before nmbd, the Squid Internet Object Cache was the process that
used more memory. That's understandable but I it just impossible for
nmbd to use all that memory.
After I saw the memory requirement of nmbd I restarted the Samba
daemons and ran top again. Here's the output:
8:32am up 34 days, 8:08h, 1 user, load average: 0.08, 0.04, 0.01
58 processes: 56 sleeping, 1 running, 0 zombie, 1 stopped
CPU states: 2.7% user, 2.7% system, 0.0% nice, 94.5% idle
USER PID %CPU %MEM NI VSZ RSS SHRD TT STAT TIME COMMAND
proxy 30619 0.0 17.2 0 14408 10912 548 ? S 1:07 squid -D -s -f /
[...]
root 11098 0.0 1.0 0 1144 660 548 ? S 0:00 /usr/sbin/nmbd -
[...]
root 11101 0.0 1.0 0 1280 648 432 ? S 0:00 /usr/sbin/smbd -
root 11100 0.0 0.8 0 1108 560 484 ? S 0:00 /usr/sbin/nmbd -
[...]
You see? Now nmbd is back to a reasonable amount of used memory
(1144+1108 KBytes).
I have two other Linux boxes running the same version of Samba in
production. I checked them and the amount of memory nmbd is using in
those system is normal. The difference is that the WINS server is the
box with nmbd using a lot of memory.
Any comments? Is it me or there is something really weird going on
here? I'll keep my eyes open to see if the situation continues.
Regards,
E.-
--
Eloy A. Paris
Information Technology Department
Rockwell Automation de Venezuela
Telephone: +58-2-9432311 Fax: +58-2-9431645