Came across this in the logs...
Nov 3 16:08:00 bubba dovecot: imap(dmiller@amfes.com): Error: Attachment file /var/mail/attachments/dc/73/dc7398c85dd02efe8a14fe6cc019b2cf07eec600-d5ca962aaae7d14c587400003bc41c5f size mismatch: 122626 != 165655
There's about a dozen different file entries listed in the error log.
I'm using 2.0.6, mdbox, and the mails are stored on a local XFS partition. I did recently start running xfs_fsr to defragment - could that do anything?
Daniel
On 4.11.2010, at 2.09, Daniel L. Miller wrote:
Nov 3 16:08:00 bubba dovecot: imap(dmiller@amfes.com): Error: Attachment file /var/mail/attachments/dc/73/dc7398c85dd02efe8a14fe6cc019b2cf07eec600-d5ca962aaae7d14c587400003bc41c5f size mismatch: 122626 != 165655
What's the file's size? What are the X metadata entries in m.* files about this? What kind of content is inside it (base64 or decoded)?
grep dc7398c85dd02efe8a14fe6cc019b2cf07eec600-d5ca962aaae7d14c587400003bc41c5f mdbox/storage/m.*
On 11/3/2010 7:13 PM, Timo Sirainen wrote:
On 4.11.2010, at 2.09, Daniel L. Miller wrote:
Nov 3 16:08:00 bubba dovecot: imap(dmiller@amfes.com): Error: Attachment file /var/mail/attachments/dc/73/dc7398c85dd02efe8a14fe6cc019b2cf07eec600-d5ca962aaae7d14c587400003bc41c5f size mismatch: 122626 != 165655 What's the file's size? What are the X metadata entries in m.* files about this? What kind of content is inside it (base64 or decoded)?
grep dc7398c85dd02efe8a14fe6cc019b2cf07eec600-d5ca962aaae7d14c587400003bc41c5f mdbox/storage/m.*
ls -al gives me 122626.
Result of grep command:
mdbox/storage/m.1643:X29609 165655 B76 dc/73/dc7398c85dd02efe8a14fe6cc019b2cf07eec600-d5ca962aaae7d14c587400003bc41c5f mdbox/storage/m.1646:/var/mail/attachments/dc/73/dc7398c85dd02efe8a14fe6cc019b2cf07eec600-d5ca962aaae7d14c587400003bc41c5f mdbox/storage/m.1646:/var/mail/attachments/dc/73/dc7398c85dd02efe8a14fe6cc019b2cf07eec600-d5ca962aaae7d14c587400003bc41c5f mdbox/storage/m.1646:dc7398c85dd02efe8a14fe6cc019b2cf07eec600-d5ca962aaae7d14c587400003bc41c5f
mdbox/storage/m.1646:dc7398c85dd02efe8a14fe6cc019b2cf07eec600-d5ca962aaae7d14c587400003bc41c5f
I believe that attachment is a pdf.
Daniel
On Wed, 2010-11-03 at 19:17 -0700, Daniel L. Miller wrote:
Nov 3 16:08:00 bubba dovecot: imap(dmiller@amfes.com): Error: Attachment file /var/mail/attachments/dc/73/dc7398c85dd02efe8a14fe6cc019b2cf07eec600-d5ca962aaae7d14c587400003bc41c5f size mismatch: 122626 != 165655
OK, I see now where there's a bug, but I can't figure out how to reproduce it. Maybe there would be another better way to fix this than I'm thinking. Could you insert a crash to this error message (attached) and get gdb backtrace from it? http://dovecot.org/bugreport.html
On Thu, 2010-11-04 at 14:01 +0000, Timo Sirainen wrote:
On Wed, 2010-11-03 at 19:17 -0700, Daniel L. Miller wrote:
Nov 3 16:08:00 bubba dovecot: imap(dmiller@amfes.com): Error: Attachment file /var/mail/attachments/dc/73/dc7398c85dd02efe8a14fe6cc019b2cf07eec600-d5ca962aaae7d14c587400003bc41c5f size mismatch: 122626 != 165655
OK, I see now where there's a bug, but I can't figure out how to reproduce it. Maybe there would be another better way to fix this than I'm thinking. Could you insert a crash to this error message (attached) and get gdb backtrace from it? http://dovecot.org/bugreport.html
Nah, forget it. I just managed to do it myself. :)
On Thu, 2010-11-04 at 14:02 +0000, Timo Sirainen wrote:
Nov 3 16:08:00 bubba dovecot: imap(dmiller@amfes.com): Error: Attachment file /var/mail/attachments/dc/73/dc7398c85dd02efe8a14fe6cc019b2cf07eec600-d5ca962aaae7d14c587400003bc41c5f size mismatch: 122626 != 165655
OK, fixed: http://hg.dovecot.org/dovecot-2.0/rev/29c71b6e354e
Everything was actually working correctly. It was just logging those errors unnecessarily.
On 11/4/2010 7:08 AM, Timo Sirainen wrote:
On Thu, 2010-11-04 at 14:02 +0000, Timo Sirainen wrote:
Nov 3 16:08:00 bubba dovecot: imap(dmiller@amfes.com): Error: Attachment file /var/mail/attachments/dc/73/dc7398c85dd02efe8a14fe6cc019b2cf07eec600-d5ca962aaae7d14c587400003bc41c5f size mismatch: 122626 != 165655 OK, fixed: http://hg.dovecot.org/dovecot-2.0/rev/29c71b6e354e
Everything was actually working correctly. It was just logging those errors unnecessarily.
Good to know nothing was actually being lost :). Thanks - looking forward to the next version.
-- Daniel
participants (2)
-
Daniel L. Miller
-
Timo Sirainen