Having a firewall in place can be an issue - the firewall may drop idle connections silently (we see this with oracle, mod_jk/ajp/memcache and in some instances cached mysql connections; unfortunately neither end know the connection has been dropped and they still send packets down the broken connection.. - so any attempt to connect just hangs even doing a database ping!
We have dropped mod_jk in favour of using mod_proxy/mod_proxy_http as this doesn't have the same problem (but is slightly less efficietnt)
You can have some success with playing the the tcp keep alive settings: echo 600 > /proc/sys/net/ipv4/tcp_keepalive_time echo 60 > /proc/sys/net/ipv4/tcp_keepalive_intvl echo 10 > /proc/sys/net/ipv4/tcp_keepalive_probesBut even these don't always seem to work - our more robust applications keep a track of when the connection was last used and will not re-connect to a connection which hasn't been used for 15 minutes
(using mod_perl this can happen!) On 17/06/2016 15:30, Joe Muller wrote:
I am working on a project to migrate all our IPlanet 6.1 SP19 webserver proxies (formerly Sun One) to Apache 2.4, since IPlanet 6.1 does not support TLS 1.2 and IPlanet 7.0 is being EOL. Our backend application servers are Weblogic 9.2 / Weblogic 12c. The IPlanet proxies have performed FLAWLESSLY for over 10 years, despite the product being no longer supported and their WL Plug-in not officially supported with Weblogic 12c. However now that we are trying to use a more supported configuration (self-compiled Apache 2.4.18 running Weblogic Server Plugin 12.1.3) we are constantly seeing these errors, which results in performance degradation for our applications, and in some cases I think maybe even lost data. [Tue Jun 14 09:27:36.239682 2016] [weblogic:error] [pid 12513:tid 140185150932736] [client 10.165.254.1:28171] <1251314659108487> Write to the client failed: calling URL::close at line 559 of BaseProxy.cpp, referer: https://intgalf.xyz.com/ALFA/selectFileType.do?fileType=O1MM [Tue Jun 14 09:27:36.239747 2016] [weblogic:error] [pid 12513:tid 140185150932736] [client 10.165.254.1:28171] <1251314659108487> ** *****Exception type [WRITE_ERROR_TO_CLIENT] raised at line 560 of BaseProxy.cpp, referer: https://intgalf.xyz.com/ALFA/selectFileType.do?fileType=O1MM [Tue Jun 14 09:27:36.239952 2016] [weblogic:error] [pid 12513:tid 140185150932736] [client 10.165.254.1:28171] <1251314659108487> request [/ALFA/servlet/DecryptDownload?linkName=al_o1mm_carr20150630.csv] did NOT process successfully.................., referer:https://intgalf.xyz.com/ALFA/selectFileType.do?fileType=O1MM Our topology is like this: Client Browser <--> Firewall <--> Load Balancer <--> Web Proxies <--> Firewall <--> Weblogic Application Servers Oracle support suggested as work around that we increase WLSocketsTimeOut in the plug-in, but I think that only masks the issue, as we still the errors. We did a network trace and it looks like the Apache plug-in is pre-maturely closing the connection to the WL server, but I can't be certain. We know that our firewall is not responsible. Any ideas ? I thought Apache would work better then Sun One, but this has been the opposite. Is there some fundamental webserver tunable parameter that is so different between out of the box Sun One and out of the box Apache that could be causing this ? --------------------------------------------------------------------- To unsubscribe, e-mail: users-unsubscribe@xxxxxxxxxxxxxxxx For additional commands, e-mail: users-help@xxxxxxxxxxxxxxxx
--The Wellcome Trust Sanger Institute is operated by Genome Research Limited, a charity registered in England with number 1021457 and a company registered in England with number 2742969, whose registered office is 215 Euston Road, London, NW1 2BE.
--------------------------------------------------------------------- To unsubscribe, e-mail: users-unsubscribe@xxxxxxxxxxxxxxxx For additional commands, e-mail: users-help@xxxxxxxxxxxxxxxx
|