search

stevenkang / spymemcached

Automatically exported from code.google.com/p/spymemcached
0 stars 0 forks source link

Spymemcached fail reconnect to local memcahed under high load ... #173

Closed GoogleCodeExporter closed 9 years ago

GoogleCodeExporter commented 9 years ago 
Hello.
My environment:
==============
OS: RHEL 5.
uname -a
Linux node1 2.6.18-164.el5PAE #1 SMP Tue Aug 18 15:59:11 EDT 2009 i686 athlon 
i386 GNU/Linux
---------
memcached version: 1.4.5
memcached options:
memcached -d -p 11211 -u memcached -c 1024 -m 3072 -t 30
nginx: 0.8.4
spymemcached: 2.5
==============
Nginx use memcached too.
Today, we restart JBOSS 4.2 AS under load.
And spymemcached started OK.
But after 15 minutes connection to memcached broken.
LOG:
---------
08:22:24 WARN  Closing, and reopening {QA sa=/10.197.162.14:11211, #Rops=3, 
#Wops=0, #iq=2, 
topRop=net.spy.memcached.protocol.binary.OptimizedGetImpl@972e040, topWop=null, 
toWrite=0, interested=1}, attempt 0.
08:22:24 WARN  Discarding partially completed op: 
net.spy.memcached.protocol.binary.OptimizedGetImpl@972e040
08:22:24 WARN  Discarding partially completed op: 
net.spy.memcached.protocol.binary.OptimizedSetImpl@972e181
08:22:24 WARN  Discarding partially completed op: 
net.spy.memcached.protocol.binary.MultiGetOperationImpl@a2dae92
08:22:24 ERROR Get objectS from memcached failed! Timeout: 5 seconds expired. 
Probably memcached is down or unavailable!java.lang.RuntimeException: Cancelled
08:22:24 WARN  Closing, and reopening {QA sa=/10.197.162.24:11211, #Rops=2, 
#Wops=0, #iq=0, 
topRop=net.spy.memcached.protocol.binary.OptimizedSetImpl@973b988, topWop=null, 
toWrite=0, interested=1}, attempt 0.

---------

Q: Why spymemcached lose connection to local memcached ?

Tell me more...

Original issue reported on code.google.com by JohnBa...@gmail.com on 20 Apr 2011 at 7:46

GoogleCodeExporter commented 9 years ago 
unfortunately, there's not enough information here to identify why.  

it can happen if you receive a large number of timeouts in a row.  By default, 
after 1000 timeouts all connections are dropped and re-established in 2.5.

Original comment by ingen...@gmail.com on 4 May 2011 at 10:52