|
Problem IT38689
|
Status: Closed |
LOOPING "WARNING: MT_AIO_WAIT: ERRNO == EINVAL OPERATION = 3" MESSAGE WHEN ADDING RAW CHUNK WHEN SIZE+OFFSET IS TOO LARGE
|
| product: |
INFORMIX SERVER / 5725A3900 / C10 - IDS 12.10 |
| Problem description: |
When adding a raw chunk and specifying a size+offset that is
larger than the raw device, it is considered expected behavior
to see one iteration of this message in the onconfig file:
WARNING: mt_aio_wait: errno == EINVAL operation = 3
Assert Warning: WARNING: mt_aio_wait: errno == EINVAL
IBM Informix Dynamic Server Version 12.10.FC14
Who: Session(, informix@host, ...)
Thread(, onspace, ...)
File: aio_gskit.c Line: 3235
See Also: /path/af....
WARNING: mt_aio_wait: errno == EINVAL
However, in certain configurations, the chunk addition can loop
for an extended period of time with many of the above messages
in MSGLOG. The stack for this scenario resembles:
(from 12.10.FC14 on Linux)
0x0000000001410459 (oninit) afstack
0x0000000001413431 (oninit) afhandler
0x0000000001413b52 (oninit) afwarn_interface
0x00000000013f19c7 (oninit) mt_aio_wait
0x00000000013f42d2 (oninit) mt_aio_start
0x00000000013f46dd (oninit) mt_aio_read
0x0000000000e767ec (oninit) verchunkx
0x0000000000e6933b (oninit) chkstat_be
0x0000000000e69964 (oninit) chkfields
0x0000000000e6bdc2 (oninit) add_chunk
0x0000000000e6fa0f (oninit) onspace
0x000000000154290b (oninit) spawn_thread
0x0000000001401c70 (oninit) th_init_initgls
0x0000000001448997 (oninit) startup
It seems as if this only impacts 12.10.xC13+ and 14.10.xC2+. |
| Problem Summary: |
****************************************************************
* USERS AFFECTED: *
* Users of Informix Server prior to 12.10.xC16 and 14.10.xC8. *
****************************************************************
* PROBLEM DESCRIPTION: *
* See Error Description *
****************************************************************
* RECOMMENDATION: *
* Upgrade to Informix Server 12.10.xC16 or 14.10.xC8. *
**************************************************************** |
| Local Fix: |
Possible workarounds:
1. Move socket poll threads to NET VPs.
2. Set onconfig parameter DIRECT_IO to 1. This may still
produce more than one error.
3. Note the value of AFWARN from onstat -g ras (likely 0x401),
run onmode -wm AFWARN=0x800, add the raw chunk(s), then set
AFWARN
back to the original value with onmode -wm AFWARN=0x401. |
| Solution |
|
| Workaround |
****************************************************************
* USERS AFFECTED: *
* Users of Informix Server prior to 12.10.xC16 and 14.10.xC8. *
****************************************************************
* PROBLEM DESCRIPTION: *
* See Error Description *
****************************************************************
* RECOMMENDATION: *
* Upgrade to Informix Server 12.10.xC16 or 14.10.xC8. *
**************************************************************** |
| Comment |
Fixed in Informix Server 12.10.xC16 and 14.10.xC8. |
| Timestamps |
Date - problem reported :
Date - problem closed :
Date - last modified :
| 12.10.2021
07.04.2022
02.05.2023 |
| Problem solved at the following versions (IBM BugInfos) |
|
| Problem solved according to the fixlist(s) of the following version(s) |
