Backup failed

[admin1979]

Hello,

We have a UNIX system Compaq Tru64 UNIX V5.1A.


We have a working backup script which gave no problems ever. But recently the backup is failed giving following error messages,

Statement just before the error occured was ,

/usr/bin/rsh remoteserver1 /tmp/backup/script.scr

1. The system has no more ptys. Ask your system administrator to create more.
2. The system only has a finite number of ptys and you have many of them in use. The usual reason for this is that you forgot (or didn't know) to call "wait" after closing each of them.

So it looks like the system is falling short of the /dev/pty ...atleast I think so. Dont know the real reason. Could anyone give some hints? Let me know if you need anymore info.

Thanx,
admin

[Vladimir Fabecic]

So looks like server1 had no more ptys during the backup time.
Log on as root on console and check out.
Create more ptys or kill some existing sessions.

[admin1979]

sorry?? Do you mean the remoteserver1 is falling short of pty?? Ohh I thought the backup server is.
Please clarify.


Thanx,
Admin

[Venkatesh BL]

what is the value of 'nptys' tunable on the system? may be it is very low.

[admin1979]

Hello,

I dont think its low,

root@backupserver > sysconfig -q pts
pts:
nptys = 255

Like you , even I felt that it needs to be increased. So I did this on the backupserver

> sysconfig -r pts nptys=400
nptys: reconfigured

> sysconfig -q pts
pts:
nptys = 400

Do you think its the right thing to do ?
We can verify the backup only tomorrow. So any other suggestions ??

[Vladimir Fabecic]

How do you run backup script, from terminal or from cron?
Did you check how many pty devices are created?

[admin1979]

We run backup from cron. how do I check how many ptys are created ...I did not get it.

Currently its ,

root@> ls -l /dev/pty* | wc -l
736

Let me know.

[Pieter 't Hart]

>>> /usr/bin/rsh remoteserver1 /tmp/backup/script.scr <<<
the backupserver seems to have copied a backup script to remoteserver1.
and tries to execute this using rsh.
rsh tries to establish a session with remoteserver1, and will use a pty on remoteserver1, not on backupserver.

so >>> root@backupserver > sysconfig -q pts
<<< most be repeated on remoteserver1!
but only gives info about configured ptys.

also check how many pty's are used on remoteserver1 "ps -ef |grep pty |wc -l"

[admin1979]

Hello,

Then I was completely wrong when I thought other way.

Anyways but the observations are same even on Remoteserver1,

# sysconfig -q pts
pts:
nptys = 255

Remoteserver1>
# ps -ef | grep pts
root 28129 15783 0.0 10:53:10 ttyr3 0:00.01 grep pts

Remoteserver1>
# ps -ef | grep pty
root 3996 15783 0.0 10:53:14 ttyr3 0:00.00 grep pty

So now I wonder should I increase this value or not. Becauze 255 seems ok like. But we never know.
Plz suggest.

Thanx,
admin

[Pieter 't Hart]

ps gives you a current situation.
it may have been different at the time of the backup failure.
can you rerun the backup?

check /var/adm/messages if anything logged at time of backup.
also check if a previous backup-process is still active.