how to fix a corrupt utmp in rhel3u4

[Matt Shaffer_1]

every other day a rhel3u4 box has an issue with utmp. we either have to get all users off the system and >utmp or reboot the server which is actually faster. is there an way to fix the utmp without having to reboot? I have found docs on fwtmp for HPUX but can't find any fixes for rhel. we have a case open with HP but wanted to open a forum as well. thanks

Matt

[Steven E. Protter · Accepted solution]

Shalom Matt,

Patch the box to update 6 or above and the problem will go away.

SEP

[Matt Shaffer_1]

Steven,

I know you've helped me out a lot in the past so I definite respect your suggestion so don't take this the wrong way: Are you serious? If I upgrade to U6 this will fix the issue? We have 400+ ml350s running rhel3u4 and this is the only server that consistently has this problem. One other server has had the problem once or twice but it hasn't happened in a long time.

thanks for the info
Matt

[Avijit Patra]

Is the box running applications (like abinitio) that uses utmp extensively? The application(s) are possibly not writing properly to the utmp. utmp become corrupted repeatedly when users/application does not close their sessions properly.

If this is the case, working with the application vendor can give result.

[Matt Shaffer_1]

upgrading the telnet-server daemon seems to have fixed the issue.