slocate hanging and unable to be killed

[Aviewanew]

The computer is an AMD 2500+ with about 9 or 10 hard drives (storage for video editing).

I had a script that ran hourly on a cron job to update a database (slocate -U /DRIVES/ -o /root/slocatedb/contentDB). Sometimes it would work fine, othertimes slocate would hang. I wouldn't always notice right away, so I might check the server and find 48+ instances of slocate running, all hung.

And I couldn't kill them! kill -9 pid, killall slocate nothing would work. I tried starting at the top, the bottom, in between. I tried killing every individual instance and still no dice.

Eventually I turned off the cronjob. Now I have another cronjob that runs every night around midnight that does an eix-sync and an slocate -u. I never had any problems with the ordinary slocate hanging - but now it has. I have two slocate -u's running, unable to be killed.

I've tried re-emerging slocate, but nothing helps. I imagine it has to do with my multitude of hard drives, cause I can't find any results online. I ran smart tests on them, and they all seem fine, they're all running, and I can get the data off them fine.

Any suggestions?

[frostschutz]

slocate does not do anything special. all it does is read in lists of files into its database. so it walks the directory tree, that is all.

if this process hangs it usually points to a hardware problem, if they cannot be killed either the process may have locked up somewhere in the kernel.

is there anything in dmesg?
does lsof or similar show any open files by the hanging processes, and are these files actually readable?
is your /etc/updatedb intact so it won't try to build a database on unreliable / temporary fs types?

another program that walks through the whole directory tree is find. does find hang as well?

In that case it would definitively be defective hardware or corrupt file system.