-
Title
Post too slow, status always "WAITING" or Reader Slow -
Description
Post too slow for all transactions and all tables, less than 100 messages per second. This applies to other shareplex processes such as Reader . Similar diagnostics need to be done to identify the issue.
1.show post detail shows "Post state" always Waiting.
sp_ctrl (splx-vip:2300)> show post detailHost : splx-vip
Source : o.SPLX_SOURCE Queue : splx-vipOperations
Target Status Posted Since Total Backlog
---------- --------------- ---------- ------------------ ---------- ----------
o.GSDLYX Running 18916 31-Mar-11 15:10:32 0 169936897Last operation posted:
Redo log: 31333 Log offset: 361200816
COMMIT "SGPM_OUT"."S_WKS_DEAL" at 03/26/11 22:09:25Post state : Waiting
Activation Id : 15
Last operation transaction id : 6
Number of open transactions : 4
Number of active transactions : 0
Number of messages read released : 12595944Number of threads : 4
Number of sessions : 6
Peak number of sessions : 6
Number of Oracle connections : 7
Concurrency (Active sessions) : 0
Number of pending messages : 0Operations posted : 18916
Transactions posted : 32Full rollbacks : 0
Full rollback operations posted : 0
Rollback operations skipped : 0Insert operations : 8326
Update operations : 95
Delete operations : 10462Insert batch operations / average : 249 / 33 rows
Update batch operations /2. read is slow when capture is on the same pace with Oracle redoRead timing debug shows high number for queue reads, which indicates it is spending time outside of Shareplex. Most likely it is related to select call bug if customer is HP-UX 11.31 IA -
Cause
HP-UX Defect,
Description:
PHKL_38700:
( QX:QXCR1000589939 SR:8606493324 CR:JAGag45497 )
When select(2m) is used as a simple timer without specifying
any file descriptors, it behaves like a timer. Currently it
is not making use of the high resolution timer functionality
already provided by the timer subsystem. Because of this
small timeouts will experience a larger wait time.
PHKL_41967:
( QX:QXCR1001118815 )
Applications that uses select(2) system call with zero timeout
argument may either suffer slower performance or return incorrectly
with EINTR on signal interruption.
-
Resolution
Resolution 1
Enabling " hires_timeout_enable" kernel parameter.
Resolution 2
Apply the patch PHKL_41967 or PHKL_38700
-
Attachments
-
Additional Information
For hp-ux 11.31, set it directly.
For hp-ux 11.23, may need install patch first.
Resolution Details:
select(2m) code has been modified to use the high resolution
timer functionality. However, the new functionality depends
on the value of tunable "hires_timeout_enable". With the
enablement of the tunable "hires_timeout_enable", along with
this patch, applications that use select(2m) as pure timer
for small timeouts will timeout right after the specified
duration and this will reduce the long wait time when compared prior to the installation of this patch.