after successful compilation nfs-utils fails to merge to the root with the following message: >>> /usr/lib/nfs/rmtab --- /usr/lib/nfs/sm/ >>> /usr/lib/nfs/sm/.keep_net-fs_nfs-utils-0 --- /usr/lib/nfs/sm.bak/ >>> /usr/lib/nfs/sm.bak/.keep_net-fs_nfs-utils-0 !!! Failed to chown/chmod/unlink in movefile() !!! /usr/lib/nfs/state !!! signed integer is greater than maximum For more details see the URL provided. Reproducible: Always Steps to Reproduce: 1. Try to merge nfs-utils with the command "emerge nfs-utils ". 2. Happens automatically with every combination of use flags. Actual Results: >>> /usr/lib/nfs/rmtab --- /usr/lib/nfs/sm/ >>> /usr/lib/nfs/sm/.keep_net-fs_nfs-utils-0 --- /usr/lib/nfs/sm.bak/ >>> /usr/lib/nfs/sm.bak/.keep_net-fs_nfs-utils-0 !!! Failed to chown/chmod/unlink in movefile() !!! /usr/lib/nfs/state !!! signed integer is greater than maximum srvkoz08 ~ # date && ls -l /usr/lib/nfs/sm.bak/.keep_net-fs_nfs-utils-0 /usr/lib/nfs/state Wed Jul 25 13:59:20 CEST 2007 -rw-r--r-- 1 root root 0 Jul 25 13:58 /usr/lib/nfs/sm.bak/.keep_net-fs_nfs-utils-0 -rw-rw-rw- 1 root root 0 Jul 25 13:58 /usr/lib/nfs/state Expected Results: >>> Regenerating /etc/ld.so.cache... portage tree is up to date and there are no packages need updating. not even with --newuse.
This is my first gentoo bugreport so i'm sorry for this weird description :) Thnaks in advance
emerge --info please.
Created attachment 125970 [details] output of: emerge --info Sorry, here it is :)
Created attachment 126200 [details, diff] raise the exception for debugging purposes What filesystem type do you have for /usr/lib/nfs/? This patch will give some more information about the exact failure. If you save it as /tmp/debug.patch then it can be applied as follows: patch /usr/lib/portage/pym/portage.py < /tmp/debug.patch Please post the output that is shown if you try to install nfs-utils with that patch applied.
Created attachment 126210 [details] merge with patch applied
> What filesystem type do you have for /usr/lib/nfs/? It is an ext3. mkfs'd with default options.
(In reply to comment #5) > os.chown(dest,sstat[stat.ST_UID],sstat[stat.ST_GID]) > OverflowError: signed integer is greater than maximum It looks like this python bug: https://sourceforge.net/tracker/?func=detail&atid=105470&aid=1747858&group_id=5470 In your case, the call that's triggering it is chown('/usr/lib/nfs/state', nobody, root). Do you have the "nobody" user listed in /etc/passwd with uid 65534?
> It looks like this python bug: > In your case, the call that's triggering it is chown('/usr/lib/nfs/state', > nobody, root). Do you have the "nobody" user listed in /etc/passwd with uid > 65534? Ah! That must be the problem as nobody has uid 4294967294. nobody:x:4294967294:65534:nobody:/:/bin/false I guess this is a default uid on 64bit systems as I didn't change it.
(In reply to comment #8) > I guess this is a default uid on 64bit systems as I didn't change it. That doesn't look normal. I suggest that you change it to 65534. Something like this will change the uid on all of the files for you: find / -mount -uid 4294967294 -print0 | xargs -0 chown 65534
the /etc/passwd and /etc/group files are exactly the same for all architectures and "nobody" has always been 65534
I guess we'll just wait for this to get fixed in python upstream.
(In reply to comment #9) > (In reply to comment #8) > > I guess this is a default uid on 64bit systems as I didn't change it. > > That doesn't look normal. I suggest that you change it to 65534. Something > like this will change the uid on all of the files for you: > > find / -mount -uid 4294967294 -print0 | xargs -0 chown 65534 > Yes, changing the nobody uid to 65534 solved the problem. Thanks guys: >>> net-fs/nfs-utils-1.0.12 merged.
