>
> > O.K. this was dev package. Obviously urpmi tried to stat every file
in
> > it - triggering devfsd into all sort of actions ... what is worse, I
> > tried rpm -Fvh - and it did the same, and even worse, it appeared to
> > *loop* contantly statting the same files in /dev.
> >
> > Strange enough, I tried to put dev into skip.list but urpmi behaved
the
> > same.
> >
> > Something really needs to be done about it ... I feared something
> > similar.
>
> Did you get the same with a basic rpm -Uvh invocation ?
>
No, I have not tried -U. I went into cooker RPMS, and obviously I would
not want rpm -U * in this case. But wait ...
[root@cooker RPMS]# rpm -Uvh dev-3.2-3mdk.i586.rpm
Preparing...
root 3532 3435 13 17:05 pts/1 00:00:09 rpm -Uvh
dev-3.2-3mdk.i586.rpm
[root@cooker root]# strace -p 3532
stat64("/dev/cciss", 0xbfffdec0) = -1 ENOENT (No such file or
directory)
stat64("/dev/ida", 0xbfffdec0) = -1 ENOENT (No such file or
directory)
stat64("/dev/cciss", 0xbfffdec0) = -1 ENOENT (No such file or
directory)
stat64("/dev/ida", 0xbfffdec0) = -1 ENOENT (No such file or
directory)
stat64("/dev/cciss", 0xbfffdec0) = -1 ENOENT (No such file or
directory)
stat64("/dev/ida", 0xbfffdec0) = -1 ENOENT (No such file or
directory)
stat64("/dev/cciss", 0xbfffdec0) = -1 ENOENT (No such file or
directory)
stat64("/dev/ida", 0xbfffdec0) = -1 ENOENT (No such file or
directory)
stat64("/dev/cciss", 0xbfffdec0) = -1 ENOENT (No such file or
directory)
stat64("/dev/ida", 0xbfffdec0) = -1 ENOENT (No such file or
directory)
stat64("/dev/cciss", 0xbfffdec0) = -1 ENOENT (No such file or
directory)
stat64("/dev/ida", 0xbfffdec0) = -1 ENOENT (No such file or
directory)
stat64("/dev/cciss", 0xbfffdec0) = -1 ENOENT (No such file or
directory)
stat64("/dev/ida", 0xbfffdec0) = -1 ENOENT (No such file or
directory)
stat64("/dev/cciss", 0xbfffdec0) = -1 ENOENT (No such file or
directory)
stat64("/dev/ida", <unfinished ...>
> Do you get a dev package to be upgrade ? this could be the cause.
>
Sure! That was I meant - updating the dev in presence of devfs would be
a problem. It turned out to be much worse than I suspected.
-andrej
