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

samba-bugs at samba.org samba-bugs at samba.org
Tue Sep 20 09:39:40 MDT 2011


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

--- Comment #10 from Eric Shubert <ejs at shubes.net> 2011-09-20 15:39:39 UTC ---
(In reply to comment #9)
> 
> I'm going to put a problem/test file on the target machine's hdd and continue
> testing with a local2local transfer.

I'm holding off on the local test for the moment. After closer look at the
strace, the problem appears to be with the transport/ssh mechanism (input side
on the target).

Early on in the strace, after a series (10x) of 4092-byte reads, I see this:
1636  07:12:20.682841 select(1, [0], [], NULL, {60, 0}) = 1 (in [0], left {60,
0})
1636  07:12:20.683211 read(0, "\374\17\0\7", 4) = 4
1636  07:12:20.683964 select(1, [0], [], NULL, {60, 0}) = 1 (in [0], left {60,
0})
1636  07:12:20.684226 read(0,
"\222E\23\272R\224<\256\0I`\365\177\247\204\207Jf\370\277=\\q\24q\357\202\"z\307\207H"...,
4092) = 4092
1636  07:12:20.684653 select(1, [0], [], NULL, {60, 0}) = 1 (in [0], left {60,
0})
1636  07:12:20.685065 read(0, "\374\17\0\7", 4) = 4
1636  07:12:20.685291 select(1, [0], [], NULL, {60, 0}) = 1 (in [0], left {60,
0})
1636  07:12:20.685627 read(0,
"v5z\260\341\346\260\335w\22\311P\37|%:2N\2177\304\0264\215#\344\274P^\307u\313"...,
4092) = 1223
1636  07:12:20.685984 select(1, [0], [], NULL, {60, 0} <unfinished ...>
1635  07:12:20.686201 <... getdents64 resumed> /* 2 entries */, 32768) = 48
1635  07:12:20.687710 getdents64(0, /* 0 entries */, 32768) = 0
1635  07:12:20.687953 close(0)          = 0

This looks like the same situation as the timeout, except that there is
"getdents64 resumed" instead of "select resumed" after the select statement.
Don't know if it's related or not.

I'll keep looking to see what I can tell from the strace.

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


More information about the rsync mailing list