URL:
  <http://savannah.cern.ch/bugs/?81584>

                 Summary: XrdCnsd looses connection and stop exporting log
files.
                 Project: XROOTD
            Submitted by: abh
            Submitted on: 2011-04-27 22:35
                Severity: 4 - Important
                Priority: 7 - High
                  Status: None
                 Privacy: Public
             Assigned to: abh
        Originator Email: 
             Open/Closed: Open
         Discussion Lock: Any
      Fixed by commit(s): 

    _______________________________________________________

Details:

Here is a history of the issue...


Thanks Alex for shedding more light on this. We’ve been having problem off
and on and now it looks like the real underlying problem is the way XrdCnsd
uses the xrootd client to copy files into the redirector. One idea we’ve
been considering is to get rid of that part altogther and just let you
specify how the copy-in is to be done. That way, you can use you favorite
program (scp, xrdcp, cp to nfs, etc). This would solve the underlying
problem. Additionally, if you choose anything other than xrdcp you would not
need to run an extra xrootd on the redirector node.
 
Andy
 
From: Alex Bogert 
Sent: Tuesday, April 26, 2011 11:32 AM
Subject: Xrootd CNSD bugged
 
Hi Everyone, 
 
I recently found an error in our xrootd setup. The CNS daemons are no longer
reporting correctly to the redirector's inventory. When I noticed files
missing in the Inventory I rebooted the xrootd system, and saw some strange
behavior. Each of the worker nodes connect's to the redirector and begins
updating their inventory, but after about 100 - 200 updates, the local
XrdCNSd reports a trashed connection to the redirector. Also, If i disable
xrootdfs about 100 extra updates will get through before the connection is
trashed. I am stumped as to the cause of this problem. I've attached the logs
from the redirector and from one worker node after a restart. 
 
I would appreciate any advice on how to resolve this.
 
Cheers,
Alex
 
 
Current xroot processes running on redirector:
 
$ ps aux | grep xroot
xrootd    8975  0.0  0.0  45424  4104 ?        Sl   Apr22   0:00
/opt/osg-v1.2/xrootd/bin//xrootd -l
/opt/osg-v1.2/xrootd/var/logs/xrdlog.atlas01.ucsc.edu -c
/opt/osg-v1.2/xrootd/etc/xrootd.cfg
xrootd    8976  0.0  0.0  41084  3856 ?        Sl   Apr22   0:00
/opt/osg-v1.2/xrootd/bin//xrootd -n cns -l
/opt/osg-v1.2/xrootd/var/logs/xrdlog.atlas01.ucsc.edu -c
/opt/osg-v1.2/xrootd/etc/xrootd.cfg
xrootd    9060  0.0  0.0 121032  4236 ?        Sl   Apr22   0:00
/opt/osg-v1.2/xrootd/bin//cmsd -l
/opt/osg-v1.2/xrootd/var/logs/cmslog.atlas01.ucsc.edu -c
/opt/osg-v1.2/xrootd/etc/xrootd.cfg
xrootd    9206  0.0  0.0 368432 12228 ?        Ssl  Apr22   0:01
/opt/osg-v1.2/xrootdfs/bin/xrootdfsd /xrootdfs/atlas -o
allow_other,fsname=xrootdfs,max_write=131072,attr_timeout=10,entry_timeout=10
 
Current xroot processing on worker (other workers look the same).
 
$ ps aux | grep xroot
1060     14469  0.0  0.0  61148   760 pts/1    S+   11:19   0:00 grep xroot
xrootd   19697  0.0  0.0  49684  4412 ?        Sl   Apr22   0:00
/opt/osg-v1.2/xrootd/bin//xrootd -l
/opt/osg-v1.2/xrootd/var/logs/xrdlog.wrk2prv.ucsc.edu -c
/opt/osg-v1.2/xrootd/etc/xrootd.cfg
xrootd   19716  0.0  0.0 123740 12672 ?        Sl   Apr22   0:00
/opt/osg-v1.2/xrootd/bin/XrdCnsd -d -D 2 -i 90 -b
atlas01.ucsc.edu:1095:/atlas/inventory
xrootd   19778  0.0  0.0  46020  3864 ?        Sl   Apr22   0:00
/opt/osg-v1.2/xrootd/bin//cmsd -l
/opt/osg-v1.2/xrootd/var/logs/cmslog.wrk2prv.ucsc.edu -c
/opt/osg-v1.2/xrootd/etc/xrootd.cfg





    _______________________________________________________

Reply to this item at:

  <http://savannah.cern.ch/bugs/?81584>

_______________________________________________
  Message sent via/by LCG Savannah
  http://savannah.cern.ch/