Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

I'm running a few dozen Debian Wheezy diskless fat clients and, while trying to diagnose a slow performance web browser issue, I noticed IceWeasel(version 24.0-1~bpo70+1) was getting a ton of recvfrom errors in strace;

3977 is IceWeasel's PID

strace -p 3977 -e 'recvfrom'

recvfrom(4, 0x7f621ffd3074, 4096, 0, 0, 0) = -1 EAGAIN (Resource temporarily unavailable)
recvfrom(4, 0x7f621ffd3074, 4096, 0, 0, 0) = -1 EAGAIN (Resource temporarily unavailable)
recvfrom(4, 0x7f621ffd3074, 4096, 0, 0, 0) = -1 EAGAIN (Resource temporarily unavailable)

etc

Upon closer inspection, it looks like they're all erroring out, so it seems to be not so temporary;

strace -c -p 3977 -e 'recvfrom'
Process 3977 attached - interrupt to quit
^CProcess 3977 detached
% time     seconds  usecs/call     calls    errors syscall
------ ----------- ----------- --------- --------- ----------------
100.00    0.001544           0      3588      3588 recvfrom
------ ----------- ----------- --------- --------- ----------------
100.00    0.001544                  3588      3588 total

I"m not terribly familiar with chasing down issues using strace so I was wondering how I might proceed here? How do I find out what file/FIFO is erroring out and why?

I thought perhaps lsof -p 3977 | grep '7f621ffd3074' would lead me somewhere but no joy.

The Fat clients are all completely diskless LTSP nodes. The root file system is mounted via NFSv3 and the /home directory via NFSv4 so it's pretty likely this is NFS related I think.

share|improve this question
    
looks like polling with UDP. Can you mount with TCP? –  tigran Jan 20 at 10:31

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.