Informix - Problem description
Problem IT13854 | Status: Closed |
-768 INTERNAL ERROR IN ROUTINE OLAP_TEMPTAB_WRITE: NO FREE BUFFE R W/ QUERY USING OLAP FUNCTIONS AND USING KAIO TO TEMP DBSPACES | |
product: | |
INFORMIX SERVER / 5725A3900 / C10 - IDS 12.10 | |
Problem description: | |
If you have a query that uses OLAP functions that have to write data to temp tables created in your temporary dbspaces, and the system is configured to use kaio then the writes to these temporary tables will fail due to improperly aligned buffers, causing the query to fail with just a -768 error. To use kaio for your temporary dbspaces, the temporary dbspace must be in a raw or block special device (so not a file system file and using kaio by enabling DIRECT_IO...in these cases, the server will do kaio to regular dbspaces, but IO to temporary dbspaces, created with the -t flag, will use aio instead). Stack if you set a trap for the 768 error: sqerr sqerr1 olap_pbyoby_next sort_open sort_open filltemp scan_open materialize_viewtmp materialize_viewtmp materialize_viewtmp materialize_viewtmp materialize_viewtmp materialize_viewtmp materialize_viewtmp prepselect open_cursor sql_open sq_open sqmain spawn_thread th_init_initgls startup The frame of interest is olap_pbyoby_next, which shows the query is doing some OLAP function...unfortunately the olap_temptab_write() function already returns before the -768 error is reported so you can not see it on the stack. From onstat -g ath you can see kaio threads, which indicate the server is using KAIO: Threads: tid tcb rstcb prty status vp-class name 32 4b291400 0 3 IO Idle 1cpu* kaio 38 4b439560 0 3 IO Idle 8cpu* kaio 39 4b4398a0 0 3 IO Idle 10cpu* kaio Then output from lsof (on Linux) which verifies the temporary dbspace is using a block special device, so kaio would be used to write to that device: COMMAND PID USER FD TYPE DEVICE SIZE/OFF NODE NAME oninit 12246 root 256u BLK 253,19 0t0 6581 /dev/mapper/vg01-lvol12 oninit 12246 root 257u BLK 253,25 0t0 6976 /dev/mapper/vg01-lvol19 oninit 12246 root 258u BLK 253,20 0t0 6645 /dev/mapper/vg01-lvol13 oninit 12246 root 259u BLK 253,21 0t0 6698 /dev/mapper/vg01-lvol14 oninit 12246 root 260u BLK 253,22 0t0 6762 /dev/mapper/vg01-lvol15 oninit 12246 root 261u BLK 253,26 0t0 7040 /dev/mapper/vg01-lvol20 oninit 12246 root 262u BLK 253,6 0t0 5563 /dev/mapper/temp-lvol1 | |
Problem Summary: | |
**************************************************************** * USERS AFFECTED: * * All users * **************************************************************** * PROBLEM DESCRIPTION: * * See Error Description * **************************************************************** * RECOMMENDATION: * * Update to IBM Informix Server 12.10.xC7 * **************************************************************** | |
Local Fix: | |
Solution | |
Workaround | |
not known / see Local fix | |
Timestamps | |
Date - problem reported : Date - problem closed : Date - last modified : | 18.02.2016 28.06.2016 28.06.2016 |
Problem solved at the following versions (IBM BugInfos) | |
12.10.xC7 | |
Problem solved according to the fixlist(s) of the following version(s) |