It's all about the answers!

Ask a question
QuestionsTagsUsersBadges
Follow this question
Question details

×42,790

question asked: Feb 23 '11, 5:23 p.m.
question was seen: 6,301 times
last updated: Feb 23 '11, 5:23 p.m.

Related questions

FAQ - How this Forum works

Problems when exposing RTC through Apache Web Server Proxy

0
Adil Chahid (455●2●41●18) | asked Feb 23 '11, 5:23 p.m.
Hi all,
it's been a while we've setup an Apache Web Server 2.2 to expose with mod_proxy our RTC Instance.
Everything worked fine for a while.
But these last days, we start seeing strange behaviours were people cannot save WI from the RTC Client nor upload change sets.
Normally those occur for a shot period and everything start working again after that.
During that period here is what we see in the Apache Web Server Error Log 

[Wed Feb 23 16:55:18 2011] [error] (113)No route to host: proxy: HTTPS: attempt to connect to 192.168.1.10:443 (rtc.mycompany.com) failed

[Wed Feb 23 16:55:18 2011] [error] ap_proxy_connect_backend disabling worker for (rtc.mycompany.com)

[Wed Feb 23 16:55:18 2011] [error] proxy: HTTPS: disabled connection for (rtc.mycompany.com)

[Wed Feb 23 16:55:18 2011] [error] proxy: HTTPS: disabled connection for (rtc.mycompany.com)



...


It seems that WebSphere Application Server invalidates the pooled proxy connection (from AWS) and when the connection Pool is maxed out, we get a 503 Error.
the rest of the exposed sites work fine so we tend to thing that something happens at the WebSphere level.

Does anyone has any clue onwhich Apache mod_proxy parameters or WAS parameters we should start looking at?

Thanks!

2 answers

Most liked answers ↑|Newest answers|Oldest answers

0
permanent link
Adil Chahid (455●2●41●18) | answered Feb 24 '11, 5:00 p.m.
Hi all,

I'm investigating if this can be related to a Possible timeout for mod_proxy
What do you think that we should use as timeouts parameter for the Proxy?
Is 30 seconds enough for the worker to initiate, use and close the socket? 


<Proxy>

   ProxySet connectiontimeout=5 timeout=30 

</Proxy>



We've also decreased the TCP Keep Alive for the Linux from 7200 to 600 seconds as follow 

net.ipv4.tcp_keepalive_time = 60



Please advise.
Thanks!
0
permanent link
Mircea Vutcovici (6) | answered Feb 28 '11, 9:50 a.m.
As a workaround you can try: 

ProxyPass .... retry=0 ttl=30

or: 

<Proxy>

   ProxySet ... retry=0 ttl=30

</Proxy>

Your answer

Register or to post your answer.


Dashboards and work items are no longer publicly available, so some links may be invalid. We now provide similar information through other means. Learn more here.