[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Hanging "svn log" caused "bdb PANIC"

From: Ulrich Eckhardt <eckhardt_at_satorlaser.com>
Date: Mon, 1 Mar 2010 10:12:39 +0100

Hi!

We had the following error message from our cronjob (which does a backup of a
dump file) this weekend:

> svnadmin: Berkeley DB error for filesystem
> '/home/svnserve/repos/development/db' while creating cursor for
> reading a string: DB_RUNRECOVERY: Fatal error, run database recovery
> svnadmin: bdb: PANIC: fatal region error detected; run recovery
> svnadmin: bdb: PANIC: fatal region error detected; run recovery

This problem has been resolved since, but I'm wondering if we didn't trigger a
bug in SVN or BDB there. I'm also anxious if the problem has just vanished
but is still lurking somewhere to nibble our behinds in the most inconvenient
time.

What are the circumstances?

 * The system uses Subversion 1.5.1 from Debian on x86.
 * Setup is several BDB repositories, the format file reports version 5.
 * Access is exclusively via svnserve, of course with the exception of the
backup task.
 * The backup is directly from the repositories, without an intermediate
hotcopy.
 * There were two hung svnserve processes[2], both hung for two weeks. These
alone didn't cause the error though, but I'm not sure if they had any
influence.

What was the immediate cause?

My colleague was trying to extract the log from the repository and parse it
for some info, using a client on win32. He used a second process connected
with a pipe and parsed the output of "svn log"[1]. The mistake he made was
that some receive buffer was too small (or never flushed), so svn couldn't
write all output it wanted to and subsequently was blocked. This in turn
seems to have caused the above frightening error message. After just killing
the parser process, without any manual intervention the backup process
finished without error next night.

Now, what I'm wondering is, is this a false alarm or should we watch the
system more closely? Is there perhaps anything we could or should do? Also,
what caused the

Thanks for your time!

Uli

[1] Yes, he knows that there's an API, but "the XML output is just too easy to
parse" according to him.

[2] Hung svnserve processes seem to occur now and then, but never cause any
harm. Still, I'd consider them bugs in svnserve. Is there anything I could do
in order to hunt these down?

-- 
ML: http://subversion.tigris.org/mailing-list-guidelines.html
FAQ: http://subversion.tigris.org/faq.html
Docs: http://svnbook.red-bean.com/
Sator Laser GmbH, Fangdieckstraße 75a, 22547 Hamburg, Deutschland
Geschäftsführer: Thorsten Föcking, Amtsgericht Hamburg HR B62 932
**************************************************************************************
Sator Laser GmbH, Fangdieckstraße 75a, 22547 Hamburg, Deutschland
Geschäftsführer: Thorsten Föcking, Amtsgericht Hamburg HR B62 932
**************************************************************************************
           Visit our website at <http://www.satorlaser.de/>
**************************************************************************************
Diese E-Mail einschließlich sämtlicher Anhänge ist nur für den Adressaten bestimmt und kann vertrauliche Informationen enthalten. Bitte benachrichtigen Sie den Absender umgehend, falls Sie nicht der beabsichtigte Empfänger sein sollten. Die E-Mail ist in diesem Fall zu löschen und darf weder gelesen, weitergeleitet, veröffentlicht oder anderweitig benutzt werden.
E-Mails können durch Dritte gelesen werden und Viren sowie nichtautorisierte Änderungen enthalten. Sator Laser GmbH ist für diese Folgen nicht verantwortlich.
**************************************************************************************
Received on 2010-03-01 10:13:20 CET

This is an archived mail posted to the Subversion Users mailing list.

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.