DO NOT REPLY [Bug 7757] New: with big file, rsync times out out when it should not; the sender is still responsive

[samba-bugs at samba.org]

https://bugzilla.samba.org/show_bug.cgi?id=7757

           Summary: with big file, rsync times out out when it should not;
                    the sender is still responsive
           Product: rsync
           Version: 3.0.7
          Platform: Other
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P3
         Component: core
        AssignedTo: wayned at samba.org
        ReportedBy: tim.liim at alcatel-lucent.com
         QAContact: rsync-qa at samba.org


Description of problem:
    When sending big file (eg. > 500MB) With --timeout=xx option, the
    server side of rsync timeouts out when it should not.  The components
    of rsync (generator or receiver or server) should exchange periodic
    keepalive msg, so they will not timeout when the connection is
    perfectly healthy.

Version-Release number of selected component (if applicable):
    rsync-3.0.7-3.fc13.i686

How reproducible:
    With --timeout=10: sometimes with 500MB file; almost always with 2GB file.

Steps to Reproduce:
    1. Assume timeout=10 sec.  generate the source file, eg. 
           i=/tmp/t55
           time dd if=/dev/zero of=$i bs=1M count=1000
       If it takes less than <timeout> x 3 (eg. 30 sec), increase
       the file size.  Bigger file better illustrates this issue.

    2. j=/tmp/t56
       rsync --timeout=10 localhost:$i $j

Actual results:
    Saw the following errors, depending on the mode of failure.

    #1
    [sender] io timeout after 10 seconds -- exiting
    rsync error: timeout in data send/receive (code 30) at io.c(140) 
           [sender=3.0.7]

    #2
    rsync: writefd_unbuffered failed to write 6 bytes to socket [generator]: 
           Broken pipe (32)
    rsync error: timeout in data send/receive (code 30) at io.c(1530) 
           [generator=3.0.7]
    rsync error: received SIGUSR1 (code 19) at main.c(1288) [receiver=3.0.7]


Expected results:
    The sync should complete without error; certainly no timeout.

Additional info:
    1. The issue is more prounced with two hosts of different speed,
       eg. h1 (dest) is much slower than h2 (src) (eg. h1 could be an
       slower machine (especially I/O bound) with LUKS; this could be
       simulated by strace on h1:receiver).
            i=/tmp/t55 j=/tmp/t56
            h2: dd if=/dev/zero      of=$i bs=1M count=1000
            h1: rsync --timeout 10   h2:$i $j

    2. From strace, the scenario with error msg in #2:
       - generator <g> read <dest> and sends checksum to 
         sender <s> (--server in this case).
       - <g> spawns receiver <r>
       - <r> opens <dest> and <tmp> file, start copying from <dest> to <tmp>
       - <s> reads <src> file and sends msgs to <r>
       - when <s> finishes, it starts to wait, for (timeout/2) sec.
       - Assume <r> is slower, <s> timeouts out after <timeout> sec.

       - <g> probably needs to send keepalive msg to <s>.

    3. strace for #1 is similar to #2: 
       - after <s> finishes, <r> takes more than 10 sec to finish 
         copying <dest> to <tmp>  and then rename <tmp> <dest>, 
         so <s> times out.

    4. The issue is somewhat reproducible on one single host.  But
       with strace on both <s> and <r>, it becomes harder to
       reproduce, probably because strace slows both <s> and <r> down.
       I need to use a much bigger file (2.5GB), and strace <r> from
       beginning, but strace <s> only torwards the end, to show this
       error with strace.

    5. There are several existing bugs that may be related to this
       unexpected timeout issue.

bug2783
    Random high loads during syncs (server side) / client stream errors
    rsync: connection unexpectedly closed (2761332 bytes received so far)
    [generator]
    rsync error: error in rsync protocol data stream (code 12) at io.c(365)

bug5478
    rsync: writefd_unbuffered failed to write 4092 bytes [sender]: 
                Broken pipe (32)

    rsync: writefd_unbuffered failed to write 4092 bytes [sender]: 
                Broken pipe (32)
    io timeout after 30 seconds -- exiting
    rsync error: timeout in data send/receive (code 30) at io.c(239) 
                [sender=3.0.2]

bug5695
    improve keep-alive code to handle long-running directory scans

    ./io.c:void maybe_send_keepalive(void)

bug6175
    write last transfer status when timeout or other error happens
    rsync: writefd_unbuffered failed to write 4 bytes [sender]: 
           Broken pipe (32)
    rsync: connection unexpectedly closed (99113 bytes received so 
           far) [sender]
    rsync error: unexplained error (code 255) at io.c(600) [sender=3.0.5]

bug7195
    timeout reached while sending checksums for very large files


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