Stephen K. Gielda
2004-Jun-06 01:03 UTC
[Dovecot] Errors and periodic corruption using 99.10.5 release with mbox
Hi, I'm seeing these errors in our logs periodically: file istream.c: line 135 (i_stream_skip): assertion failed: (stream->v_size == 0 || stream->v_offset + count <= stream->v_size) Timeout leak: 0x8069328 Any idea on what may be causing these? System is freebsd, format is mbox. I'm also seeing periodic corruption of mailboxes. It's easily fixed because it's always a few stray lines of text left at the beginning of the mbox file above the From I'm currently trying to find out more about what actions may cause it. /steve
James Moser
2004-Jun-06 01:20 UTC
[Dovecot] Errors and periodic corruption using 99.10.5 release with mbox
I've seen the same thing... in .4 we had blank lines at the top of the file, which always seemed to fix themselves... however with .5 you get an error stating the file is not in mbox format and it dies. There are usually three lines at the top, with only a few characters in the second line: (blank) somegarbage here (blank) From blah at doh.com ..... The error messages with .4 were: Jun 5 00:06:23 hostname pop3(username): Error indexing mbox file /var/mail/username: LF not found where expected But with .5, they simply state the file is not in mbox format and fails. Stephen K. Gielda wrote:> Hi, > > I'm seeing these errors in our logs periodically: > > file istream.c: line 135 (i_stream_skip): assertion failed: > (stream->v_size == 0 || stream->v_offset + count <= stream->v_size) > > Timeout leak: 0x8069328 > > Any idea on what may be causing these? System is freebsd, format is > mbox. > > I'm also seeing periodic corruption of mailboxes. It's easily fixed > because it's always a few stray lines of text left at the beginning of > the mbox file above the From I'm currently trying to find out more > about what actions may cause it. > > /steve-- James L Moser james at powweb.com PowWeb Hosting http://www.powweb.com /(bb|[^b]{2})/, that is the Question. mysql>SELECT * FROM user WHERE clue > 0; Empty set (0.03 sec) Health is merely the slowest possible rate at which one can die... Health nuts are going to feel stupid someday, lying in hospitals dying of nothing... -------------- next part -------------- A non-text attachment was scrubbed... Name: james.vcf Type: text/x-vcard Size: 239 bytes Desc: not available URL: <http://dovecot.org/pipermail/dovecot/attachments/20040605/30ecd3be/attachment-0001.vcf>
Stephen K. Gielda
2004-Jun-06 01:31 UTC
[Dovecot] Errors and periodic corruption using 99.10.5 release with mbox
At 09:20 PM 6/5/2004, James Moser wrote:>I've seen the same thing... in .4 we had blank lines at the top of the >file, which always seemed to fix themselves... however with .5 you get an >error stating the file is not in mbox format and it dies. There are >usually three lines at the top, with only a few characters in the second line: > >(blank) > somegarbage here >(blank) > From blah at doh.com ..... > >The error messages with .4 were: > >Jun 5 00:06:23 hostname pop3(username): Error indexing mbox file >/var/mail/username: LF not found where expected > >But with .5, they simply state the file is not in mbox format and fails.This is exactly what I've seen too in regards to the corruption (both .5 and .4). /steve>Stephen K. Gielda wrote: > >>Hi, >> >>I'm seeing these errors in our logs periodically: >> >>file istream.c: line 135 (i_stream_skip): assertion failed: >>(stream->v_size == 0 || stream->v_offset + count <= stream->v_size) >> >>Timeout leak: 0x8069328 >> >>Any idea on what may be causing these? System is freebsd, format is mbox. >> >>I'm also seeing periodic corruption of mailboxes. It's easily fixed >>because it's always a few stray lines of text left at the beginning of >>the mbox file above the From I'm currently trying to find out more about >>what actions may cause it. >> >>/steve > > >-- >James L Moser james at powweb.com >PowWeb Hosting http://www.powweb.com > >/(bb|[^b]{2})/, that is the Question. > >mysql>SELECT * FROM user WHERE clue > 0; >Empty set (0.03 sec) >Health is merely the slowest possible rate at which one can die... Health >nuts are going to feel stupid someday, lying in hospitals dying of nothing... > > >