dmesg returns proc: table is full

[Gary Mc Adam]

When I run dmesg it returns 200+ lines of:

proc: table is full

Any ideas?

Jason

[John Poff]

Hi,

Check your /var/adm/syslog/syslog.log file and see if you have any recent complaints there. The dmesg program reads from a ring buffer which wraps around. Your messages could be old. If they are new, you should check your nproc kernel parameter. It might need to be bumped up.

JP

[Jason Hawkins]

I found these lines in the syslog.log file, but the only other thing before this was a failed rcp command:

Aug 12 13:49:50 novec_a vmunix: proc: table is full
Aug 12 13:50:08 novec_a vmunix: proc: table is full
Aug 12 13:50:21 novec_a above message repeats 6739 times

Also nproc is set to 1024 and the system is at about 160 processes right now.

Jason

[Tor-Arne Nostdal]

Check your kernel parameters !

Have you tried to search the forum for this message and read previous answers?

I did a fast search and found several links, and here's two ;-)

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=98099

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=686615

/Tor-Arne

[Tor-Arne Nostdal]

Hi again..

As John said - it could be old messages...
The syslog messages is from Aug 12

If you have a correct date setting that should be quite some time ago.

Check only the tail of the syslog file.

/Tor-Arne

[KCS_1]

Hi,Jason

it seems like old message or temporary the peak value of the days.

Can you see the message today in syslog??

If you can't, have a look at the follow


# man dmesg

/var/adm/messages (message buffer) shows the contents when you type 'dmesg' command.

[Nick Tarmey_3]

Hi,

I had this issue a few weeks back:

Aug 11 14:35:01 hpserv01 vmunix: file: table is full
Aug 11 14:35:01 hpserv01 above message repeats 87 times

So I increased the maxusers default of 32 to 64 and I haven't seen the problem since.

Changing this value had a knock on effect on the following params values shown before and after.

maxfiles 32 64
ksi_alloc_max 2208 4256
ncallout 292 548
nclist 612 1124
ncsize 1500 1788
nfile 910 1371
ninode 476 764
nkthread 499 947
nproc 276 532

Of course your kernel will need to be rebuilt and the system rebooted.

Good luck!
Nick

[Kent Ostby]

Gary --

A message may appear in dmesg output that no longer applies to the system since dmesg is just a circular repository of messages.

Do :

grep "table is full" /var/adm/syslog/syslog.log | tail -20

This will show you when the last 20 of these happened.

If you are concerned about it reoccurring, you can increase nproc or maxusers to increase the size of the proc table.

Best regards,

Kent M. Ostby

[bhavin asokan]

hi,

increase the value of nproc ,nfile,maxusers and max_thread_proc

regds,