tomcat-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From bugzi...@apache.org
Subject DO NOT REPLY [Bug 44454] - busy count reported in mod_jk inflated, causes incorrect balancing
Date Tue, 19 Feb 2008 23:50:35 GMT
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG·
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://issues.apache.org/bugzilla/show_bug.cgi?id=44454>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND·
INSERTED IN THE BUG DATABASE.

http://issues.apache.org/bugzilla/show_bug.cgi?id=44454





------- Additional Comments From anoonan@indeed.com  2008-02-19 15:50 -------
     Well, we're a busy site, but not stunningly so... everyone's idea on what
makes a busy site may differ.  The httpd systems are never that terribly loaded
CPU/Load wise (less then 40%).  The global counter and the individual counters
add up to each over, so I believe that indicates that they both are having the
issue.  It's hard to tell at what level the inflation occurs, as without calming
servers at around the same time each day it's hard to tell what are real
requests and which are inflation.  We have a few servers in rotation that are
bulkier and run a 2x higher LBFactor, and they seem currently to have about 2x
the false busys.  I'm guessing about 5 a day for the lighter servers, and 10 a
day for the bigger servers.  Because the load balancing is unequal, it varies,
but since our graceful restart 2.5 days ago, the ACC on the servers is anywhere
from 2-4 million, depending on the server.
     As for the balancing method, the busyness method is definitely more even
when it works since our requests can be very different weights.  As for running
1.2.26, I think we could compile/use it no problem (we're running 2.2.3, the
precompile indicates 2.2.6 as a prereq), but I know management would be a little
grumbly to make a change without a better idea as to what's going on.  I don't
think the config changes are a problem for us as we use preference files and
JKMounts only at VHost level... no globals.
     I'm gathering the gstack data now and will post in a few minutes.

In case it comes up, as typical worker is configured like:
worker.<WORKER>.port=8009
worker.<WORKER>.host=<IP>
worker.<WORKER>.type=ajp13
worker.<WORKER>.socket_timeout=30
worker.<WORKER>.connection_pool_timeout=60
worker.<WORKER>.lbfactor=20

The LB involved is:
worker.<LB>.lock=P
worker.<LB>.method=B

and we have a global setting of:
worker.maintain=15

Thanks,
Andrew

-- 
Configure bugmail: http://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@tomcat.apache.org
For additional commands, e-mail: dev-help@tomcat.apache.org


Mime
View raw message