PQ72333: FAILOVER WITH PROXY DOES NOT WORK.

 Fixes are available

5.0.1: WebSphere Application Server Version 5.0 Fix Pack 1 (Version 5.0.1)
5.0.1: WebSphere Application Server Enterprise Edition Version 5.0 Fix Pack 1



APAR status
Closed as program error.

Error description
Customer test with pack dump in the balance port proxy and
plug-in does making connection to second available clone but do
not send any request to second clone. As well the warning in the
logs says that it failed to get the request parameter.
[Thu Dec 26 13:10:52 2002] 00006079 00000016 - TRACE:
connection:
Keep-Alive
[Thu Dec 26 13:10:52 2002] 00006079 00000016 - TRACE:
cache-control:
no-cache
[Thu Dec 26 13:10:52 2002] 00006079 00000016 - TRACE:    cookie:
wfacookie=B-184206364;
JSESSIONID=0001BNERWTOIBV5YGQJLTRWTCDA:tv4spin9
[Thu Dec 26 13:10:52 2002] 00006079 00000016 - TRACE:    $WSIS:
false
[Thu Dec 26 13:10:52 2002] 00006079 00000016 - TRACE:    $WSSC:
http
[Thu Dec 26 13:10:52 2002] 00006079 00000016 - TRACE:    $WSPR:
HTTP/1.1
[Thu Dec 26 13:10:52 2002] 00006079 00000016 - TRACE:    $WSRA:
10.18.14.166
[Thu Dec 26 13:10:52 2002] 00006079 00000016 - TRACE:    $WSRH:
10.18.14.166
[Thu Dec 26 13:10:52 2002] 00006079 00000016 - TRACE:    $WSSN:
wlsz128.win.wellsfargo.com
[Thu Dec 26 13:10:52 2002] 00006079 00000016 - TRACE:    $WSSP:
8000
[Thu Dec 26 13:10:52 2002] 00006079 00000016 - TRACE:
lib_htrequest:
htrequestWrite: Writing the request content
[Thu Dec 26 13:10:52 2002] 00006079 00000016 - TRACE:
lib_htrequest:
htrequestWrite: content length is 1305
[Thu Dec 26 13:10:52 2002] 00006079 00000016 - TRACE:
lib_htrequest:
htrequesWrite: Reading the POST content cache interested = 0
[Thu Dec 26 13:10:52 2002] 00006079 00000016 - TRACE: ns_plugin:
cb_read_body: In the read body callback
[Thu Dec 26 13:11:22 2002] 00006079 00000016 - WARNING:
ns_plugin:
cb_read_body: Failed to read the request body
[Thu Dec 26 13:11:22 2002] 00006079 00000016 - TRACE: ws_common:
websphereExecute: Wrote the request; reading the response
[Thu Dec 26 13:11:22 2002] 00006079 00000016 - TRACE:
lib_htresponse:
htresponseRead: Reading the response:
[Thu Dec 26 13:11:22 2002] 00006079 00000016 - TRACE:
lib_htresponse:
htresponseSetError: Setting the error |1|
[Thu Dec 26 13:11:22 2002] 00006079 00000016 - ERROR: ws_common:
websphereExecute: Failed to read from a new stream; App Server
may have
gone down during read
[Thu Dec 26 13:11:22 2002] 00006079 00000016 - TRACE:
lib_stream:
destroyStream: Destroying the stream
[Thu Dec 26 13:11:22 2002] 00006079 00000016 - ERROR: ws_server:
serverSetFailoverStatus: Marking onefile down
[Thu Dec 26 13:11:22 2002] 00006079 00000016 - ERROR: ws_common:
websphereHandleRequest: Failed to execute the transaction to
'onefile';
will try another one
[Thu Dec 26 13:11:22 2002] 00006079 00000016 - ERROR: ws_common:
websphereHandleRequest: Failed to find an app server to handle
this
request
[Thu Dec 26 13:11:22 2002] 00006079 00000016 - TRACE: ws_common:
websphereEndRequest: Ending the request
[Thu Dec 26 13:11:22 2002] 00006079 00000016 - TRACE: ws_cache:
cacheFinish: In cacheFinish
.
Error reports with Failed to read from a new stream; App Server
may have gone down during read. This type of problem was
discovered with WAS 4.0.4 environment and we provided e-Fix to
customer to resolve, but customer applied the PTF 5 and PQ66649
address the problem is included in new rollout PTF defect list.
Customer still complain about the similar problem is reported
with new PTF.
.
This to work for long term we need to add an attribute for the
server element called ExtendedHandshake similar to what was
added in 3.5. This causes the plugin to perform some extended
handshaking to ensure the connect() call was successful.
Local fix Problem summary
****************************************************************
* USERS AFFECTED: WebSphere Application Server version 4.0     *
*                 users of the webserver plugin with a proxy   *
*                 firewall in between the plugin and the       *
*                 app server.                                  *
****************************************************************
* PROBLEM DESCRIPTION: The plugin was not failing over         *
*                      as expected when a proxy firewall is    *
*                      between the plugin and the app server.  *
****************************************************************
* RECOMMENDATION:                                              *
****************************************************************
The plugin marks a server down when the connect() fails.
However, when a proxy firewall is in between the plugin and
the app server, the connect() will succeed even though the
backend app server is down.  This causes the plugin to not
failover correctly to other app servers.
Problem conclusion
The plugin now performs some handshaking with the app server
to ensure that it is up before sending the request.  This
allows us to failover in the event the app server is really
down.  This is specified using the ExtendedHandshake attribute
for the Server element.
Temporary fix Comments
APAR information
APAR number PQ72333
Reported component name WAS BASE 5.0
Reported component ID 5630A3600
Reported release 00W
Status CLOSED PER
PE NoPE
HIPER NoHIPER
Special Attention NoSpecatt
Submitted date 2003-03-21
Closed date 2003-03-21
Last modified date 2003-03-21

APAR is sysrouted FROM one or more of the following:
PQ69608

APAR is sysrouted TO one or more of the following:

Modules/Macros
PLUGIN          

Publications Referenced

Fix information
Fixed component name WAS BASE 5.0
Fixed component ID 5630A3600

Applicable component levels
R00S PSY    UP
R00A PSY    UP
R00W PSY    UP
R00H PSY    UP
R003 PSY    UP
R00I PSY    UP


Document Information


Product categories: Software > Application Servers > Distributed Application & Web Servers > WebSphere Application Server > General
Operating system(s):
Software version: 00W
Software edition:
Reference #: PQ72333
IBM Group: Software Group
Modified date: Mar 21, 2003