httpd-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From sridhar basam <sridhar.ba...@gmail.com>
Subject Re: [users@httpd] mod_proxy_ajp "dialog to [...] failed"
Date Thu, 31 May 2012 21:07:27 GMT
On Thu, May 31, 2012 at 3:08 PM, Roney Duilio Stein
<roney.stein@sondait.com.br> wrote:
> Hello.
>
> Please, I appreciate if someone could help me answering some questions:
>
> I'm working on an issue that I sent to the Tomcat users mailing list (http://tomcat.markmail.org/search/?q=#query:%20list%3Aorg.apache.tomcat.users+page:5+mid:4w7j4oqnuxwkz67c+state:results).
>
> I have some questions:
> 1. Are the error messages below caused by Apache waiting for a response from Tomcat which
never comes? There's no directly corresponding error message in the Tomcat logs, some errors
are past or before the dates on the Apache log;
> 2. After the node enters the "error state" what happens to the existing connections to
this Tomcat instance?
>
> [Thu May 31 10:15:47 2012] [error] (OS 10060)A connection attempt failed because the
connected party did not properly respond after a period of time, or established connection
failed because connected host has failed to respond.  : ajp_ilink_receive() can't receive
header
> [Thu May 31 10:15:47 2012] [error] ajp_read_header: ajp_ilink_receive failed
> [Thu May 31 10:15:47 2012] [error] (70007)The timeout specified has expired: proxy: dialog
to 127.0.0.1:9109 (127.0.0.1) failed
> [Thu May 31 10:15:49 2012] [error] proxy: BALANCER: (balancer://wlb). All workers are
in error state for route (wt2)
> [Thu May 31 10:15:50 2012] [error] proxy: BALANCER: (balancer://wlb). All workers are
in error state for route (wt2)
> [Thu May 31 10:15:51 2012] [error] proxy: BALANCER: (balancer://wlb). All workers are
in error state for route (wt2)
>
> We were suspecting on a TCP/IP stack issue (or something related) but everything works
fine (no other connection problems), only this application shows these errors. I'm suspecting
this is caused by some misbehavior in the application, which in turn, causes other problems
due to the node entering the "error state".
> Apache is proxying to 2 Tomcat instances (this application) and to 1 third application
(which doesn't suffer any problem).
>
> Apache is version 2.2.22, Tomcat 6.0.29 (same thing with 6.0.20).
> This single error puts the worker in "error state", which is bad, and the fix from 2.2.22
to overcome this kind of situation doesn't cover this specific case.
>
> The configuration used in Apache HTTP is:
> ServerSignature Off
> ServerTokens Prod
> Timeout 600
> LimitRequestFieldSize 20480
> ProxyIOBufferSize 21504
> ProxyRequests Off
> ProxyPreserveHost On
> <Proxy *>
>  Order deny,allow
>  Allow from all
> </Proxy>
> <Proxy balancer://wlb>
>  BalancerMember ajp://10.11.102.224:9109 route=wt1 loadfactor=50 max=85 ttl=120 retry=5
connectiontimeout=5000ms ping=5000ms
>  BalancerMember ajp://127.0.0.1:9109 route=wt2 loadfactor=50 max=85 ttl=120 retry=5
connectiontimeout=5000ms ping=5000ms
> </Proxy>
> ProxyPass /app1 balancer://wlb/app1 stickysession=JSESSIONID nofailover=On
> ProxyPass /app2 ajp://10.11.102.219:8009/app2
>
> The AJP connector in Tomcat (app1):
> <Connector port="9109" protocol="AJP/1.3" redirectPort="8443" packetSize="22528" maxThreads="200"
connectionTimeout="120000"/>
>

You just might not be having enough threads in this connector to
handle the workload. See how many of the 200 threads are busy when the
problem occurs. netstat -s -t would also tell you about failed
connection attempts.

 Sridhar





>
> Thanks a lot.
> Roney
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe@httpd.apache.org
> For additional commands, e-mail: users-help@httpd.apache.org
>

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@httpd.apache.org
For additional commands, e-mail: users-help@httpd.apache.org


Mime
View raw message