rsync - Dec 2009 - --timeout not honoured

Hi

I already wrote about this problem half a year ago but didn't get
an answer: http://lists.samba.org/archive/rsync/2009-June/023412.html

I'm using rsync 3.0.3 on a NAS. In the parameter list I use --timeout=1800.
But still I often have rsync stall much longer than that.

Some examples:

2009/12/14 23:14:35 [8707] io timeout after 11670 seconds -- exiting
2009/12/14 23:14:35 [8707] rsync error: timeout in data send/receive (code 30)
at io.c(239) [sender=3.0.3pre1]

2009/12/14 00:20:04 [19046] io timeout after 15485 seconds -- exiting
2009/12/14 00:20:04 [19046] rsync error: timeout in data send/receive (code 30)
at io.c(239) [sender=3.0.3pre1]

2009/12/12 22:15:21 [29732] io timeout after 8082 seconds -- exiting
2009/12/12 22:15:21 [29732] rsync error: timeout in data send/receive (code 30)
at io.c(239) [sender=3.0.3pre1]

2009/12/11 23:14:23 [30216] io timeout after 11613 seconds -- exiting
2009/12/11 23:14:23 [30216] rsync error: timeout in data send/receive (code 30)
at io.c(239) [sender=3.0.3pre1]

I know that 3.0.3pre1 is not the newest version, but as this is running
on a NAS it's not that easy to update or compile a newer version.

However, why is it that rsync waits much longer than given in --timeout?
Shouldn't timeout work that way?

Thanks

bye  Fabi

At 10:27 15.12.2009 +0100, Fabian Cenedese wrote:
>Hi
>
>I already wrote about this problem half a year ago but didn't get
>an answer: http://lists.samba.org/archive/rsync/2009-June/023412.html
>
>I'm using rsync 3.0.3 on a NAS. In the parameter list I use
--timeout=1800.
>But still I often have rsync stall much longer than that.
>
>2009/12/14 23:14:35 [8707] io timeout after 11670 seconds -- exiting
>2009/12/14 23:14:35 [8707] rsync error: timeout in data send/receive (code
30) at io.c(239) [sender=3.0.3pre1]
>
>2009/12/14 00:20:04 [19046] io timeout after 15485 seconds -- exiting
>2009/12/14 00:20:04 [19046] rsync error: timeout in data send/receive (code
30) at io.c(239) [sender=3.0.3pre1]
>
>However, why is it that rsync waits much longer than given in --timeout?
>Shouldn't timeout work that way?
I think I know now where my confusion comes from. The reported time is not
the actual time without action but the time since the start of the rsync
command.

2009/12/15 20:00:02 [4260] building file list
2009/12/15 20:00:03 [4260] .d..t...... vmware/FTPServer2/
2009/12/15 20:23:45 [4260] <f.st...... vmware/FTPServer2/FTPServer2-s017.vmdk
2009/12/15 22:12:25 [4260] <f+++++++++ vmware/FTPServer2/FTPServer2-s018.vmdk
2009/12/15 23:10:19 [4260] io timeout after 11372 seconds -- exiting
2009/12/15 23:10:19 [4260] rsync error: timeout in data send/receive (code 30)
at io.c(239) [sender=3.0.3pre1]

11372 seconds is 189.something minutes, so 20:00+3h09 makes about the 23:10
timeout time. Same seems to be true for my previous examples, all started at
20:00.

So --timeout does seem to be honoured, just the reported timeout time is a bit
misleading. If this is not to be changed maybe an explaining comment in the
docs might be a good idea.

Thanks

bye  Fabi