Follow-up Comment #1, bug #81162 (project xrootd):
Hi all,
I have come across the same error on an fresh ALICE installation
(xrootd-vmss_1.8b-1) under heavy load (the system has ~240TiB of almost free
capacity, so it receives continuous 2Gbit data stream. The error is
reproducible, but it only occurs after ~10hours of running. We don't see this
error on any other xrootd nodes (but they don't receive this load since they
are already full).
Would a gdb trace log help?
System info:
SL 5.7
xrootd-vmss_1.8b-1
kernel 2.6.18-274.12.1.el5
[root@xrootd3 server]# tail xrdlog
120228 04:39:14 1664 alicesgm.3199:[log in to unmask] oss_AioWrite: Write
262144@134217728 started; aiocb=0x1703e770
120228 04:39:14 1664 alicesgm.3199:[log in to unmask] oss_AioWait: write
completed; rc=0 result=262144 aiocb=0x1703e770
120228 04:39:14 1664 alicesgm.14778:[log in to unmask] oss_AioWrite: Write
262144@346030080 started; aiocb=0x1741d168
120228 04:39:14 1664 alicesgm.14778:[log in to unmask] oss_AioWait: write
completed; rc=0 result=262144 aiocb=0x1741d168
120228 04:39:14 1664 XrootdXeq: alicesgm.29813:[log in to unmask] disc
0:01:12 (protocol data length error)
120228 04:39:14 1664 alicesgm.29813:[log in to unmask] ftsofs_close: use=1
fn=/09/44150/51f3025c-6190-11e1-8547-ff18d5480cd7
120228 04:39:14 1664 oss_Adjust: used=5548929285658+409731072 path=/mnt/fs5/
120228 04:39:14 1664 oss_Adjust: free=29889162153984-409731072
path=/mnt/fs5/
120228 04:39:14 1664 alicesgm.29813:[log in to unmask] ftsofs_close: use=0
fn=dummy
120228 04:39:14 1664 XrdLink: client closed socket polling
alicesgm.29813:[log in to unmask]
=======================
grep -i seg /var/log/messages*:
/var/log/messages:Feb 28 04:39:14 xrootd3 kernel: xrootd[15510]: segfault at
0000000000000198 rip 000000315b608ca0 rsp 0000000056ed2e78 error 4
_______________________________________________________
Reply to this item at:
<http://savannah.cern.ch/bugs/?81162>
_______________________________________________
Message sent via/by LCG Savannah
http://savannah.cern.ch/
########################################################################
Use REPLY-ALL to reply to list
To unsubscribe from the XROOTD-DEV list, click the following link:
https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=XROOTD-DEV&A=1
|