Update of bug #100633 (project xrootd):
Assigned to: None => abh
_______________________________________________________
Follow-up Comment #3:
The only two possible solutions I can provide is
a) Increase the thread limit as the default is the number of memory pages/4.
That's pretty conservative for servers. So, do something
echo 2048 > /proc/sys/kernel/threads-max
Redhat appears not to be so so strict and defaults to a high value to avoid
such problems. For instance at SLAC on RHEL6-64 it is 125608 on RHEL5-64 it
is 32099 (I have read that even the downloadable kernel sets it to 2048). So,
I don't see a problem increasing the number for a server.
b) Setup a proxy cluster to handle the load. Client connections would then be
spread among several machines. The size should be determined by the expected
load.
As you point out, this has been an issue with the old client. We won't have
an immediate resolution until we switch to using the new client for the proxy
which should happen in the next couple of months. The new client is more
cautious about threads.
_______________________________________________________
Reply to this item at:
<http://savannah.cern.ch/bugs/?100633>
_______________________________________________
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
|