[Dovecot] timeout when retrieving folder list w/ Evolution 2.0?
hi,
i have a problem with some maildir folders, where Evolution times out when retrieving the header list. It only happens with one particular folder, but i can't seem to find a pattern behind it. The amount of headers it does manage to retrieve varies, from between 3% to 50%.
Its not a particularly large folder either (only 520 messages).
The error message I get from Evolution is "Server response ended too soon".
I did an Ethereal trace, and the closest IMAP command before the failure is:
FLAGS (\Answered \Seen) RFC822.SIZE xxx INTERNALDATE "xxx" UID nnn)\r\n
its a different UID value that fails each time though. after this response from the server, it seems to get through the message header (not sure if the entire message header was sent back).
i'll keep this folder around as a test case. any ideas on where to start debugging? i'm fairly familiar with C and gdb.
leon
On Tue, 5 Oct 2004 10:57:50 +1300, leon breedt bitserf@gmail.com wrote:
I did an Ethereal trace, and the closest IMAP command before the failure is: correction, that should be IMAP response from server.
leon
On 5.10.2004, at 00:57, leon breedt wrote:
i have a problem with some maildir folders, where Evolution times out when retrieving the header list. It only happens with one particular folder, but i can't seem to find a pattern behind it. The amount of headers it does manage to retrieve varies, from between 3% to 50%.
Its not a particularly large folder either (only 520 messages).
The error message I get from Evolution is "Server response ended too soon".
I did an Ethereal trace, and the closest IMAP command before the failure is:
FLAGS (\Answered \Seen) RFC822.SIZE xxx INTERNALDATE "xxx" UID nnn)\r\n
Nothing in server logs? Sounds like it crashed. This is with 0.99.11?
i'll keep this folder around as a test case. any ideas on where to start debugging? i'm fairly familiar with C and gdb.
If it crashed, gdb backtrace would help. http://dovecot.org/bugreport.html
On Tue, 5 Oct 2004 13:42:47 +0300, Timo Sirainen tss@iki.fi wrote:
Nothing in server logs? Sounds like it crashed. This is with 0.99.11? Nothing but the auth Login: line. This is version 1.0-test46 on RHEL3...compiled using a tweaked spec file.
It doesn't seem like it crashed though, as closing down Evolution and starting it back up makes it work...its like just that particular connection stops responding. New connections keep on working for other accounts. Don't see any sig11 messages in the logs or any other messages indicating early exit. And CPU utilization looks normal after it stops responding. Couldn't find any coredumps either..
The maildir was one I copied over from Courier IMAP, and I followed the Wiki migration instructions. I also tried shutting down Courier, rm -rf on .INBOX, and removing all the generated files I could find in the maildir, but it still occurs.
Would a complete transcript of all the IMAP commands leading up to this issue help? The service is sitting behind Perdition, so I can log that too.
Leon
On Wed, 6 Oct 2004 00:12:33 +1300, leon breedt bitserf@gmail.com wrote:
I also tried shutting down Courier, ^^^^^ I mean, Dovecot :)
On 5.10.2004, at 14:12, leon breedt wrote:
It doesn't seem like it crashed though, as closing down Evolution and starting it back up makes it work...its like just that particular connection stops responding. New connections keep on working for other accounts. Don't see any sig11 messages in the logs or any other messages indicating early exit. And CPU utilization looks normal after it stops responding. Couldn't find any coredumps either..
So the connection doesn't actually die immediately, it just doesn't do anything and evolution eventually timeouts it?
Would a complete transcript of all the IMAP commands leading up to this issue help? The service is sitting behind Perdition, so I can log that too.
Maybe. It sounds like there is still something wrong with the new nonblocking writing code..
participants (2)
-
leon breedt
-
Timo Sirainen