File Table full problem

UNIX Solaris
1

Hi All,

In my Big brother monitoring, I can see File table is 97.4% full. On searching more, I can see it is complaining for kernel open files, it seems.
If it is issue of too many open files, can we resolve it with reboot ?

# /usr/local/bin/lsof | wc -l
89132
/# /usr/local/bin/lsof | grep java | wc -l
40303
/# /usr/local/bin/lsof | grep httpsd.wo | wc -l
2971

This is Solaris-10 and was just rebooted 39 days back.
Suggest please.

Regards

2

I am guessing you mean 'file table overflow'. I am also guessing this is an older version of Solaris.

You will have to change /etc/system to bump up maxusers; reboot. The problem is too many concurrent open files.

The kernel needs to have additional open file descriptor space.

> getconf OPEN_MAX
4097
>echo "maxusers/D" | adb -k
physmem 7da8d4
maxusers:
maxusers:       2048

In this system: maxusers * OPEN_MAX == how many concurrently open files you can have.

Typically you can bump maxusers to up to: 2 * MAX_DEFAULT_MAXUSERS
-- which for the Solaris 10 box in the example above is 4096

Reboot may fix it temporarily.

3

Hi,

Thanks for replying. Here are my current settings. We rebooted it recently

/root# cat /etc/release
                      Solaris 10 10/08 s10s_u6wos_07b SPARC
           Copyright 2008 Sun Microsystems, Inc.  All Rights Reserved.
                        Use is subject to license terms.
                            Assembled 27 October 2008
/root# uname -a
SunOS tsapid01 5.10 Generic_144488-17 sun4u sparc SUNW,Netra-T12
/root# uptime
  3:10pm  up 39 day(s), 10:12,  1 user,  load average: 3.27, 3.38, 3.38
/root# getconf OPEN_MAX
256
/root# echo "maxusers/D" | adb -k
physmem fb9665
maxusers:
maxusers:       2048
/root# cat /etc/system | grep -i maxuser
*               set maxusers=40
4

Hi Jim,

I just came to know the exact picture for my problem. For file-table full problem, my Admin says to check "sar -v", Big brother is sending alert from this output. Currently Big Brother Monitoring is showing as 96.4%. I am not sure, how to check it and where it is over-flowing. Here are my details of server

root@tsapid01:/# sar -v 5 5
SunOS tsapid01 5.10 Generic_144488-17 sun4u    09/28/2011
14:32:50  proc-sz    ov  inod-sz    ov  file-sz    ov   lock-sz
14:32:55 1592/30000    0 144285/144285    0 42255/42255    0    0/0
14:33:00 1592/30000    0 144285/144285    0 42258/42258    0    0/0
14:33:05 1590/30000    0 144285/144285    0 42251/42251    0    0/0
14:33:10 1590/30000    0 144285/144285    0 42251/42251    0    0/0
14:33:15 1591/30000    0 144285/144285    0 42246/42246    0    0/0

Please help me to understand and what should be fix.

5

Can anybody help me in this please ?