Laurent Julliard writes:
> Today at about 5pm GMT Savannah became very loaded all of a sudden.
> looking at the number of running processes, there is an impressively
> high number of CVS processes running and some of them have been around
> for hours so it looks like there are problems with cvs processes hanging.
I'm investigating this.
> As a side not I noticed that /cvsroot resides on the / filesystem where
> there is only 400 MB of free space. This is not much. In addition cvs
> often uses *a lot* of temporary storage space especially for commits and
> by default it uses /tmp which also resides on the / file system.
I'm mounting a 10GB partition for /cvsroot. That will remove this
load from the / partition.
> I guess a /home/large_tmp or /audio-video/large_tmp should be created
> and then use the -T argument of cvs to specify the new cvs temp
> directory. Shouldn't the /cvsroot dir be moved to /audio-video as well
> by the way?
The problem is that there are subtle bugs when using a separate
LockDir (which is the case of all projects) on repositories with symbolic
links in them (which is why cvsroot was moved to / in the first place).
> Another thing: each time I run the top command or ps, I have a bunch of
> "Internal Error!" messages showing on the screen which is quite strange.
> is it possible that the ps and top command versions are not in sync
> with /proc
This is because ps was not recompiled with the N_GROUP_MAX fix.
It is harmless.
--
Loic Dachary http://www.dachary.org/ [EMAIL PROTECTED]
12 bd Magenta http://www.senga.org/ [EMAIL PROTECTED]
75010 Paris T: 33 1 42 45 07 97 [EMAIL PROTECTED]
GPG Public Key: http://www.dachary.org/loic/gpg.txt
