Fixes are available
6.1.0.1: WebSphere Application Server V6.1.0 Fix Pack 1 for HP-UX
6.1.0.1: WebSphere Application Server V6.1.0 Fix Pack 1 for AIX
6.1.0.1: WebSphere Application Server V6.1.0 Fix Pack 1 for Windows
6.1.0.1: WebSphere Application Server V6.1.0 Fix Pack 1 for Linux
6.1.0.1: WebSphere Application Server V6.1.0 Fix Pack 1 for i5/OS
PK25804; 6.0.2.9: after running WebSphere portal server for a couple of hours
6.1.0.1: WebSphere Application Server V6.1.0 Fix Pack 1 for Solaris
APAR status
Closed as program error.
Error description
Deadlock in Dynamic cache when processing the batch operations WebSphere Version(Base & ND) 6.0.2.9. After running portal for a couple of hours portal experience dead lock in dynacache code.This issue is seen while Stress test of Portal application on WebSphere Version 6.0.2.9.
Local fix
Test fix was provided.
Problem summary
**************************************************************** * USERS AFFECTED: WebSphere Application Sever version 6 users * * using DistributedMap. * **************************************************************** * PROBLEM DESCRIPTION: Deadlock in Dynamic cache when * * performing a batch operation. * **************************************************************** * RECOMMENDATION: * **************************************************************** The deadlock occurs under the following scenarios: 1. One thread: process clear command, it locks Cache A. It processes the batch operations. When it tries to trim the Cache B, it locks Cache B. 2. Another thread: process clear command, it locks Cache B. It processes the batch operations. When it tries to trim Cache A, it locks Cache A.
Problem conclusion
The problem is fixed by removing trim cache logic from the batch operations. The trim cache operations will be performed after a wakeup thread removes the entries whose time limit has expired. This way will prevent the deadlock problem on cache instances. The code is modified to move the process of trim cache for each cache instances in BatchUpdate daemon to TimeLimit daemon. The TimeLimit daemon is triggered when it is time to check whether the entries in cache buffer expire or not. After the code checks the expiration, it starts trimming each cache instance if buffer size of the cache instance is over the cache size. The fix for this APAR is currently targeted for inclusion in fixpack 6.0.2.13 and 6.1.0.1. Please refer to the recommended updates page for delivery information: http://www.ibm.com/support/docview.wss?rs=180&uid=swg27004980
Temporary fix
The code is modified to move the process of trim cache for each cache instances in BatchUpdate daemon to TimeLimit daemon. The deadlock occurs under the following scenarios: 1. One thread: process clear command, it locks Cache A. It processes the batch operations. When it tries to trim the Cache B, it locks Cache B. 2. Another thread: process clear command, it locks Cache B. It processes the batch operations. When it tries to trim Cache A, it locks Cache A. The TimeLimit daemon is triggered when it is time to check whether the entries in cache buffer expire or not. After the code checks the expiration, it starts trimming each cache instance if buffer size of the cache instance is over the cache size.
Comments
APAR Information
APAR number
PK25804
Reported component name
WEBS APP SERV N
Reported component ID
5724H8800
Reported release
60A
Status
CLOSED PER
PE
NoPE
HIPER
NoHIPER
Special Attention
NoSpecatt
Submitted date
2006-05-31
Closed date
2006-06-21
Last modified date
2006-06-21
APAR is sysrouted FROM one or more of the following:
APAR is sysrouted TO one or more of the following:
Modules/Macros
DYNACACH
Fix information
Fixed component name
WEBS APP SERV N
Fixed component ID
5724H8800
Applicable component levels
R60A PSY
UP
R60H PSY
UP
R60I PSY
UP
R60P PSY
UP
R60S PSY
UP
R60W PSY
UP
R60Z PSY
UP
R61A PSY
UP
R61H PSY
UP
R61I PSY
UP
R61P PSY
UP
R61S PSY
UP
R61W PSY
UP
R61Z PSY
UP
Document Information
Modified date:
18 October 2021