All,


We are getting the below exceptions on all our systems where Memcache is 
getting used. This is ultimately causing the the whole system becoming 
unresponsive .


*
*

*2012-07-14 11:00:55,841 ERROR [MaintThread] 
com.danga.MemCached.SockIOPool:1435 - ++++ failed to close SockIO obj from 
deadPool*

*2012-07-14 11:00:55,841 ERROR [MaintThread] 
com.danga.MemCached.SockIOPool:1436 - ++++ socket or its streams already 
null in trueClose call*

*java.io.IOException: ++++ socket or its streams already null in trueClose 
call*

*        at 
com.danga.MemCached.SockIOPool$SockIO.trueClose(SockIOPool.java:1704) 
~[MemCached-2.0.1.jar:na]*

*        at com.danga.MemCached.SockIOPool.selfMaint(SockIOPool.java:1432) 
~[MemCached-2.0.1.jar:na]*

*        at 
com.danga.MemCached.SockIOPool$MaintThread.run(SockIOPool.java:1497) 
[MemCached-2.0.1.jar:na]*



We are using Memcache version : 1.4.5. We are using danga client version 
2.0.1 . Below are the connection properties of Memcache


mmt.cache.initialConnections=100

mmt.cache.minSpareConnections=100

mmt.cache.maxSpareConnections=1000

mmt.cache.maxIdleTime=100000

mmt.cache.maxBusyTime=300000

mmt.cache.maintThreadSleep=5000

mmt.cache.socketTimeOut=15000

mmt.cache.socketConnectTO=3000

mmt.cache.failover=false

mmt.cache.nagleAlg=false

mmt.cache.aliveCheck=false

mmt.cache.enableCompression=false


This is causing real issues as the System on Memcache is the base of an 
Production  site.   We have looked into memcache stats and nothing looks 
unreasonable inclusive of "listen_disabled_num" which was 0. Memcache 
server Instance has 64Cores and 12Gb RAM . We are using hashing as 
NEW_COMPAT_HASH 
and have 3 servers of memcache on which data is distributed.


Any help around the same would be highly appreciated.


Regards,

Shubham


Please let us know what t look in.

Reply via email to