We are running the HP DDMi agent on our UNIX servers and it appears that it runs the 'finger -m' command.
The finger command appears to hang for 5 minutes causing a spike in CPU and then timesout.
Checking here on this site to see if there are any know issues with finger or what might cause finger -m to hang. I can run this from the command line with the same results.
The finger command only looks up usernames and login data.
How's your username lookup? Is it purely local and file-based, or are you using NIS, LDAP, Samba winbind or anything similar? (What does the "passwd:" line of your /etc/nsswitch.conf say?)
Do commands like "who" and "w" work normally? (Is your /etc/utmp[sx]? corrupted?)
Does somebody at your site have a habit of killing old sessions with "kill -9"?
How big is your /var/adm/wtmps file (or wtmpx, or wtmp, depending on HP-UX version). If it has reached a multi-gigabyte size, try making an archival copy of the old file and truncating the original to zero size. The finger command may not have been updated to handle files larger than 2 GB, and might be getting confused by the size of the file.
MK... thank you for your response... I have been doing some research and believe that this might be related to a known bug with utmpd and that patch PHCO_40529 might resolve this. We are going to test this patch to see if it is the case.
