[xml/sgml] /svn/debian-xml-sgml wedged?
Mike Hommey
mh@glandium.org
Mon, 31 May 2004 17:32:19 +0900
Jay Berkenbilt wrote:
> I did two checkouts from
>
> svn+ssh://alioth.debian.org/svn/debian-xml-sgml/
>
> that worked just fine, but now I am no longer able to do checkouts.
> The checkout just hangs.
>
> If I log into alioth and run
>
> svn ls --verbose file:///svn/debian-xml-sgml
>
> it "hangs" without generating any output. Running this command under
> strace shows what appears to be an infinite busy-wait loop:
>
> select(0, NULL, NULL, NULL, {0, 1000}) = 0 (Timeout)
> select(0, NULL, NULL, NULL, {0, 2000}) = 0 (Timeout)
> select(0, NULL, NULL, NULL, {0, 4000}) = 0 (Timeout)
> select(0, NULL, NULL, NULL, {0, 8000}) = 0 (Timeout)
> select(0, NULL, NULL, NULL, {0, 16000}) = 0 (Timeout)
> select(0, NULL, NULL, NULL, {0, 25000}) = 0 (Timeout)
> select(0, NULL, NULL, NULL, {0, 25000}) = 0 (Timeout)
> select(0, NULL, NULL, NULL, {0, 25000}) = 0 (Timeout)
> . . .
>
> I haven't investigated further. I had an svn checkout command running
> the background for over 3 hours. (I started it, went to do something
> else, came back, and found it still running.)
>
> I had to kill my svnserve processes and my svn ls processes with kill
> -KILL.
> > This smells a bit like a Berkeley db issue.
Even db4.2_deadlock does the select loop thing when looking through
strace...
> Who needs to fix this, and where do I report it?
I've used svnadmin hotcopy to make a copy of the currently fucked
repository and put it instead, but its get irremediably fucked by itself
after a while, even though it works perfectly just after the unfuckage try.
I wonder if there aren't any cron jobs doing some naughty stuff...
The error I get after a while is:
svn: Berkeley DB error while checkpointing after Berkeley DB transaction
for filesystem /svn/debian-xml-sgml/db:
Invalid argument
Now I think only the admins (Cced) can do something about it...
Mike