sshd[2424]: error: openpty: Not enough space

[Richard Pereira_1]

Customer called to say he was expenriencing lags when trying to ssh into the system, i didnt notice anything out of the ordinary except for these messages in syslog.log

May 17 09:51:07 host sshd[2424]: error: openpty: Not enough space
May 17 09:51:07 host sshd[2427]: error: session_pty_req: session 0 alloc fail
ed
May 17 09:51:07 host sshd[2427]: Disconnecting: fork failed: Not enough space
May 17 09:51:07 host sshd[2427]: clnt_dg_create: out of memory


>ps -ef | grep ssh
root 911 1 0 May 8 ? 0:00 /opt/ssh/sbin/sshd
root 2538 911 0 09:53:28 ? 0:00 sshd: psoft [priv]
hrsup 1854 1852 0 09:28:13 ? 0:00 sshd: hrsup@pts/0
psoft 2545 2538 0 09:53:34 ? 0:00 sshd: psoft@pts/1
root 1852 911 0 09:28:02 ? 0:00 sshd: hrsup [priv]
root 2680 911 0 09:57:52 ? 0:00 sshd: jcaplet [priv]
root 2970 2781 1 10:06:54 pts/3 0:00 grep ssh
root 2778 911 0 10:00:53 ? 0:00 sshd: root@pts/3
jcaplet 2682 2680 0 09:57:55 ? 0:00 sshd: jcaplet@pts/2
>who
psoft pts/ta May 17 09:53
hrsup pts/0 May 17 09:28
psoft pts/1 May 17 09:53
jcaplet pts/2 May 17 09:57
root pts/3 May 17 10:00

top shows 99% idle and Load averages: 0.02, 0.05, 0.16

>sar -v 5 5

HP-UX psa1h31 B.11.11 U 9000/800 05/17/05

10:03:16 text-sz ov proc-sz ov inod-sz ov file-sz ov
10:03:21 N/A N/A 432/8212 0 5201/67744 0 9284/67754 0
10:03:26 N/A N/A 432/8212 0 5201/67744 0 9283/67754 0
10:03:31 N/A N/A 432/8212 0 5201/67744 0 9283/67754 0
10:03:36 N/A N/A 432/8212 0 5201/67744 0 9283/67754 0
10:03:41 N/A N/A 432/8212 0 5201/67744 0 9283/67754 0

dmesg is squeaky clean


>swapinfo -tam
Mb Mb Mb PCT START/ Mb
TYPE AVAIL USED FREE USED LIMIT RESERVE PRI NAME
dev 8192 5243 2949 64% 0 - 1 /dev/vg00/lvol2
reserve - 2948 -2948
memory 3111 2237 874 72%
total 11303 10428 875 92% - 0 -


swapinfo is high, but seems stable, besides this are there any other leads??

[Steven E. Protter]

92 percent memory usage, memory defined as swap plus ram.

So its not a big leap to be out of resources.

I recommend getting more memory as the best option.

Secondary, if swap is less than twice ram, you can increase swap and not purchase ram.

Might want to run vmstat and see how hard the system is paging. glance/gpm should show high resource usage.

The best solution to high resouce use is to provide more resources.

SEP

[Thummalu]

Hi

I am also facing a similar situation. In my case the swapinfo shows 99 % full.
root% swapinfo -tam
Mb Mb Mb PCT START/ Mb
TYPE AVAIL USED FREE USED LIMIT RESERVE PRI NAME
dev 2104 1445 659 69% 0 - 1 /dev/vg00/lvol2
reserve - 659 -659
memory 1133 1106 27 98%
total 3237 3210 27 99% - 0 -
root%

Any suggestions please..

br
thummalu