[Bug 5478] rsync: writefd_unbuffered failed to write 4092 bytes [sender]: Broken pipe (32)

[samba-bugs at samba.org]

https://bugzilla.samba.org/show_bug.cgi?id=5478

Eric Shubert <ejs at shubes.net> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |ejs at shubes.net

--- Comment #7 from Eric Shubert <ejs at shubes.net> 2011-09-20 01:16:05 UTC ---
I've been running 3.0.7 setup for months with no problem, until June 2011 when
this bug showed up. I subsequently upgraded to 3.0.8, but the error persists.

Seems to be triggered by certain file content. The error happened most
frequently with files that were individual email messages (Maildir format)
containing attached pictures. When I remove the offending source file and
restart, more files are processed ok until another offending file is reached.
If I restart w/out removing the offending file, it always gets stuck on the
same file until it's removed from the list to be processed.

Then it happened with a quickbooks backup (.qbb) file as well. My guess is that
there's something in the data stream that's causing something i/o wise to seize
up.

I have kept examples of files that cause this. I set up the rsync-debug script
on the target server (not running daemon mode, just ssh), and the strace around
the error at the end is such:
31031 select(1, [0], [], NULL, {60, 0}) = 1 (in [0], left {60, 0})
31031 read(0,
"H\10\26}\23565\6\300RV\331\23\262M\357Q\210ngTQ\373\3523\326\6\36
2\246\"\1\375"..., 4092) = 4092
31031 select(1, [0], [], NULL, {60, 0}) = 1 (in [0], left {60, 0})
31031 read(0, "\374\17\0\7", 4)         = 4
31031 select(1, [0], [], NULL, {60, 0}) = 1 (in [0], left {60, 0})
31031 read(0,
"+R~\344\374I\204\344_<\311\235\307\340iT|q\32z=\373(\243\320\332\
32527x\336b"..., 4092) = 4092
31031 select(1, [0], [], NULL, {60, 0}) = 1 (in [0], left {60, 0})
31031 read(0, "\374\17\0\7", 4)         = 4
31031 select(1, [0], [], NULL, {60, 0}) = 1 (in [0], left {60, 0})
31031 read(0,
"\206/\227-W\370+\337\367\251\\h\247\346\330\233\356s\353\254\276\356\365a\206%wS\357\234:\377"...,
4092) = 4092
31031 select(1, [0], [], NULL, {60, 0}) = 1 (in [0], left {60, 0})
31031 read(0, "\374\17\0\7", 4)         = 4
31031 select(1, [0], [], NULL, {60, 0}) = 1 (in [0], left {60, 0})
31031 read(0,
"\262\231\222\225%\370\374\262.\262\21A\222V\300\336\0226\262\246\325=*\314\234\314\204Z[(\351w"...,
4092) = 4092
31031 select(1, [0], [], NULL, {60, 0} <unfinished ...>
31030 <... select resumed> )            = 0 (Timeout)
31030 select(4, [3], [], NULL, {60, 0} <unfinished ...>
31031 <... select resumed> )            = 0 (Timeout)
31031 select(1, [0], [], NULL, {60, 0} <unfinished ...>
31030 <... select resumed> )            = 0 (Timeout)
31030 select(4, [3], [], NULL, {60, 0} <unfinished ...>
31031 <... select resumed> )            = 0 (Timeout)
31031 select(1, [0], [], NULL, {60, 0} <unfinished ...>
31030 <... select resumed> )            = 0 (Timeout)
31030 select(4, [3], [], NULL, {60, 0} <unfinished ...>
31031 <... select resumed> )            = 0 (Timeout)
etc.

If there's anything I can do to help further to get this bug fixed, please let
me know.

Thanks.

-- 
Configure bugmail: https://bugzilla.samba.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.