home clear 64x64
en blue 200x116 de orange 200x116 info letter User
suche 36x36
Latest versionsfixlist
14.10.xC11 FixList
12.10.xC16.X5 FixList
11.70.xC9.XB FixList
11.50.xC9.X2 FixList
11.10.xC3.W5 FixList
Have problems? - contact us.
Register for free anmeldung-x26
Contact form kontakt-x26

Informix - Problem description

Problem IT08621 Status: Closed

READAHEAD THREAD CAN RUN CONTINOUSLY BLOCKING A CPU VP HEREBY AND RESULTING
IN HANG OF THE INSTANCE

product:
INFORMIX SERVER / 5725A3900 / C10 - IDS 12.10
Problem description:
This has been observed during a migration from 11.70 to 12.10 
with an instance that has about 3700 compressed tables. 
 
During the migration to 12.10 an attached index on the 
compression dictionary partition is created. 
 
The instance message log is showing: 
 
14:34:48  *** FromVersion=18  ToVersion=25 entity_ver=26 
targetVersion=28 indexVersion=5 *** 
14:34:48  *** FromVersion=25  ToVersion=26 entity_ver=26 
targetVersion=28 indexVersion=6 *** 
14:34:48  *** FromVersion=26  ToVersion=27 entity_ver=26 
targetVersion=28 indexVersion=7 *** 
14:34:48  *** RSAM 1 12.10.xC1 *** 
 
The convert thread is in the state ready while the readahead 
thread is continuously running on the CPU VP. There is only one 
CPU VP configured. 
 
The stack of the convert thread is showing: 
 
... 
kyadd 
rsaddindex 
rsam_1210_convert 
do_each_version_work 
generic_fw_main 
convertSM 
th_init_initgls 
startup 
 
So you are seeing in an onstat -g ath: 
 
 19       45b765f0         0                3    ready 
1cpu*        kaio 
 24       45e19028         44c2a928         3    running 
1cpu         readahead_0 
 21       45c5d808         44c28ee8         3    IO Wait 
1cpu         convert 
 
and onstat -g glo shows that the CPU time used is increasing for 
CPU VP 1.
Problem Summary:
**************************************************************** 
* USERS AFFECTED:                                              * 
* users migrating to 12.10 on a single CPU system with a       * 
* significant number of compressed tables                      * 
**************************************************************** 
* PROBLEM DESCRIPTION:                                         * 
* See Error Description                                        * 
**************************************************************** 
* RECOMMENDATION:                                              * 
* Update to IBM Informix Server 12.10.xC6                      * 
****************************************************************
Local Fix:
As a temporary relief you can attempt the migration again. 
 
- Stop current instance "kill -9 on PID of CPU VP 1". 
- Use onrestorept to revert to previous instance state before 
the migration or use a backup. 
- Change onconfig parameter AUTO_READAHEAD to 0 and VPCLASS to 
use more than one CPU VP. 
- Disable KAIO (on Linux for example set environment variable 
KAIOOFF before instance startup). 
- Start migration again with oninit -v.
Solution
Problem Fixed In IBM Informix Server 12.10.xC6
Workaround
not known / see Local fix
Timestamps
Date  - problem reported    :
Date  - problem closed      :
Date  - last modified       :
29.04.2015
30.12.2015
30.12.2015
Problem solved at the following versions (IBM BugInfos)
12.10.xC6
Problem solved according to the fixlist(s) of the following version(s)
12.10.xC6 FixList