WebSphere Application outages wrote:
> Hi,
>
> I am using Websphere 6.1 with fix pack 27 same is for IBM http server, It is observed that both the servers (WebShere Apps and HTTP) goes down frequently which is impacting end-users to process any request.
>
> on reviewing logs i got certain error message.
>
> [Tue Nov 17 15:04:02 2009] 00001e75 b75cbb90 - ERROR: /blddir/WAS61.NATV/NATV/ws/code/plugins.http/src/common/http/lib_rio.c line 884 : Read failed, rc=104
> [Tue Nov 17 15:04:02 2009] 00001e75 b75cbb90 - ERROR: ws_server: serverSetFailoverStatus: Marking was-lichfl2Node01_lickastle2 down
> [Tue Nov 17 15:04:02 2009] 00001e75 b75cbb90 - ERROR: ws_common: websphereHandleRequest: Failed to execute the transaction to 'was-lichfl2Node01_lickastle2'on host 'was-lichfl2'; will try another one
>
> can any one help to isolate the problem.
>
> Regards
> Manoj Wadhwa

You'll need to look in the appserver logs for lickastle2 to see what was
happening there at the same time. This means the HTTP server couldn't
communicate with that appserver instance, but does not necessarily mean
the request failed, it's possible that another instance successfully
processed the request.

Ken

If there is firewall between IHS and WAS then check the firewall logs to see
if connections are being dropped.

- Sunit


"WebSphere Application outages" wrote in
message
news:86740189.1258452726104.JavaMail.wassrvr@ltsgw as009.sby.ibm.com...
> Hi,
>
> I am using Websphere 6.1 with fix pack 27 same is for IBM http server, It
> is observed that both the servers (WebShere Apps and HTTP) goes down
> frequently which is impacting end-users to process any request.
>
> on reviewing logs i got certain error message.
>
> [Tue Nov 17 15:04:02 2009] 00001e75 b75cbb90 - ERROR:
> /blddir/WAS61.NATV/NATV/ws/code/plugins.http/src/common/http/lib_rio.c
> line 884 : Read failed, rc=104
> [Tue Nov 17 15:04:02 2009] 00001e75 b75cbb90 - ERROR: ws_server:
> serverSetFailoverStatus: Marking was-lichfl2Node01_lickastle2 down
> [Tue Nov 17 15:04:02 2009] 00001e75 b75cbb90 - ERROR: ws_common:
> websphereHandleRequest: Failed to execute the transaction to
> 'was-lichfl2Node01_lickastle2'on host 'was-lichfl2'; will try another one
>
> can any one help to isolate the problem.
>
> Regards
> Manoj Wadhwa

rc=104 is "connection reset by peer", meaning either the JVM crashed or some intermediate firewall/LB/router closed this connection abruptly. You'll need to corroborate thsi with logs in the AppServer or network ops.

We just had a similar problem. WAS 6.1.0.25 on Windows. IHS process went to 90% CPU with these errors in the plug-in log:
[Mon Nov 23 10:22:53 2009] 00000a60 00000f58 - ERROR: G:\blddir\WAS61.NATV\NATV\ws\code\plugins.http\src \common\http\lib_rio.c line 868 : Read failed, rc=420
[Mon Nov 23 15:11:34 2009] 00000a60 00000f4c - ERROR: G:\blddir\WAS61.NATV\NATV\ws\code\plugins.http\src \common\http\lib_rio.c line 868 : Read failed, rc=420
[Tue Nov 24 13:48:16 2009] 00000a60 00000f50 - ERROR: G:\blddir\WAS61.NATV\NATV\ws\code\plugins.http\src \common\http\lib_rio.c line 350 : Read failed, rc=420
[Tue Nov 24 14:00:28 2009] 00000a60 00000f60 - ERROR: G:\blddir\WAS61.NATV\NATV\ws\code\plugins.http\src \common\http\lib_rio.c line 350 : Read failed, rc=420