[PLUG] rsyc/scp stalls arbitrarily

[Mayuresh]

On Fri, Jul 23, 2010 at 06:55:20AM +0530, Shridhar Daithankar wrote:
> > Would appreciate any help in this regard.
> 
> strace?
> 
> -- 
> Regards 
>  Shridhar

While the transfer is on it shows read-write pairs of calls continuously.
When it stalls, following pattern keeps repeating. Seems it is polling for
something. read calls stop as write is failing and it remains in polling
loop. (This trace is taken from machine that tries to scp TO the machine
that has fault.)


Before stalling also, a few odd write calls fail but then the transfer
resumes again till it reaches a stage where there are only failed calls.

"Resource temporarily unavailable" hints at some network issue?? Not sure.


(First write that is succeeding seems to be of the status bar.)

write(1, "\ropenoffice.org-core-3.2.0-12.24"..., 85) = 85
rt_sigaction(SIGALRM, NULL, {0xb522b0, [], SA_RESTORER|SA_INTERRUPT,
0x682e08}, 8) = 0
alarm(1)                                = 0
sigreturn()                             = ? (mask now [])
write(6,
"\322\200n\265\325\236\341\r\343\7\303\362\316\f\271e\277\375\320\370<>\220\364\350w\360#\203O\331\231"...,
16384) = -1 EAGAIN (Resource temporarily unavailable)
poll([{fd=6, events=POLLOUT}], 1, -1)   = ? ERESTART_RESTARTBLOCK (To be
restarted)
--- SIGALRM (Alarm clock) @ 0 (0) ---
getpgrp()                               = 4384
ioctl(1, TIOCGPGRP, [4384])             = 0
time(NULL)                              = 1279885557
...

Mayuresh.