Not able to login to the server

[Shameer.V.A]

Hi ,

We are not able to log in to an hp-ux server using ssh.We logged in to the box from console .While we checked the number of users logged on to the box ,we can see lot users with the following status.

$ w
10:55am up 183 days, 21:17, 50 users, load average: 0.11, 0.13, 0.20
User tty login@ idle JCPU PCPU what
nrv727 console 12:34pm -sh
gpy166 ttyp1 10:11am 2 -su
iwv646 ttyp2 10:54am -sh
aze016 ttyp4 9:49am 11 -su
hqb255 ttyp5 12:59pm1557:49 -
hqb255 ttyp6 11:10am355:26 -
qnf025 ttyp7 2:28pm334:33 -
hqb255 ttyp8 2:39pm2778:36 -
dhi147 ttyp9 3:22pm1345:42 -
hqb255 ttypa 3:41pm4878:27 -
gpy166 ttypb 4:35pm4799:01 -
dhi147 ttypc 4:54pm2274:36 -
gpy166 ttypd 10:49am5717:11 -
gpy166 ttype 11:55am2063:50 -
cls375 ttypf 12:06pm42885:54 -


The total number us users currently present on the server is 54.

[likid0]

for the time being to solve the problem on a short notice, from console you can increase the number of terminals, changing:

#kctune| grep -i pty
npty 60 Default
nstrpty 60 60

check:
man npty
man nstrpty

[smatador]

Hi,
Does sshd run actually? You could stop/start ssh.

[T G Manikandan]

can you check syslog.log to see for what its complaining?

do you notice any error messages on the console.

[點燃]

This happens when users log in with ssh and does not logout properly, may due to VPN issues,
Kill those old sessions if not required,
and as a permanent fix, increase 'npty'

regards!

[Shameer.V.A]

Hi ,

the current values are


$ kmtune | grep -i pty
npty 60 - 60
nstrpty 60 - 60

Also we checked from glance and seen that the used value of npty is 6 only.

Also we cant restart the sshd daemon as this is a db production server .

[Shameer.V.A]

We are not able to kill these sessions are these are not existing

[Dennis Handly]

>We are not able to kill these sessions are these are not existing

It's possible there is a connection between a corrupted /etc/utmp file and being unable to login to your server. It may forget to free the tty and update utmp?

See this other thread about corrupted utmp:
http://forums.itrc.hp.com/service/forums/questionanswer.do?threadId=1327261