Shared memory segment message on intialisation
Posted in 2004
Topics: Performance & Tuning, Installation, Setup & Upgrades, Storage & Space Management, Server Administration, Logging & Checkpoints
7.31 UD8 on SCO Unixware 7.1.1
Our customer has just upgraded from UC6 -> UD8, and is seeing these shared
memory segment messages at start-up time. We've not seen them before on UD8
on other OSs (AIX or Linux).
has anyone else seen them, know what they mean, know if they are important
or know hopw to avoid them?
thanks
Neil
======================================================================
Wed Oct 27 13:42:50 2004
13:42:50 Event alarms enabled. ALARMPROG = '/usr/informix/etc/log_full.sh'
13:43:05 DR: DRAUTO is 0 (Off)
13:43:05 Contiguous shared memory segment allocation failed at 0xbfbc8000.Allocation retried at 0xb4400000.
Check SHMBASE is consistent with the value in $INFORMIXDIR/etc/onconfig.std.
Consider using a different SHMBASE value in your ONCONFIG file.
If shared memory segments are not allocated in increasing address order,
memory block allocation performance may degrade.
13:43:05 Contiguous shared memory segment allocation failed at 0x6c144000.Allocation retried at 0xb3c00000.
Check SHMBASE is consistent with the value in $INFORMIXDIR/etc/onconfig.std.
Consider using a different SHMBASE value in your ONCONFIG file.
If shared memory segments are not allocated in increasing address order,
memory block allocation performance may degrade.
13:43:05 Dynamically allocated new message shared memory segment (size
1616KB)
13:43:05 Contiguous shared memory segment allocation failed at 0x6c144000.Allocation retried at 0xb3400000.
Check SHMBASE is consistent with the value in $INFORMIXDIR/etc/onconfig.std.
Consider using a different SHMBASE value in your ONCONFIG file.
If shared memory segments are not allocated in increasing address order,
memory block allocation performance may degrade.
13:43:05 Dynamically allocated new message shared memory segment (size
1616KB)
13:43:05 Contiguous shared memory segment allocation failed at 0x6c144000.Allocation retried at 0xb2c00000.
Check SHMBASE is consistent with the value in $INFORMIXDIR/etc/onconfig.std.
Consider using a different SHMBASE value in your ONCONFIG file.
If shared memory segments are not allocated in increasing address order,
memory block allocation performance may degrade.
13:43:05 Dynamically allocated new message shared memory segment (size
1616KB)
13:43:06 IBM Informix Dynamic Server Version 7.31.UD8 Software SerialNumber AAC#J874023
13:43:06 IBM Informix Dynamic Server Initialized -- Shared MemoryInitialized.
13:43:06 Physical Recovery Started.
13:43:06 Physical Recovery Complete: 0 Pages Restored.
13:43:06 Logical Recovery Started.
13:43:09 Logical Recovery Complete. 0 Committed, 0 Rolled Back, 0 Open, 0 Bad Locks
13:43:10 Dataskip is now OFF for all dbspaces
13:43:10 Checkpoint Completed: duration was 0 seconds.
13:43:10 Checkpoint loguniq 622486, logpos 0x380018
13:43:10 On-Line Mode
13:43:10 Affinitied VP 5 to phys proc 5
13:43:11 Affinitied VP 1 to phys proc 2
13:43:11 Affinitied VP 4 to phys proc 4
13:43:11 Affinitied VP 7 to phys proc 7
13:43:11 Affinitied VP 6 to phys proc 6
13:43:11 Affinitied VP 3 to phys proc 3
13:48:39 Checkpoint Completed: duration was 0 seconds.1
I shoul add that I've checked the SHMBASE and stune kernel parameters for
the server and they all accord with the release notes.
"Neil Truby" <neil.truby@ardenta.com> wrote in message
news:2uaih3F28uitrU1@uni-berlin.de...
> 7.31 UD8 on SCO Unixware 7.1.1
>
> Our customer has just upgraded from UC6 -> UD8, and is seeing these shared
> memory segment messages at start-up time. We've not seen them before on
> UD8 on other OSs (AIX or Linux).
>
> has anyone else seen them, know what they mean, know if they are important
> or know hopw to avoid them?
>
> thanks
> Neil
> ======================================================================
> Wed Oct 27 13:42:50 2004
>
> 13:42:50 Event alarms enabled. ALARMPROG =
> '/usr/informix/etc/log_full.sh'
> 13:43:05 DR: DRAUTO is 0 (Off)
> 13:43:05 Contiguous shared memory segment allocation failed at> 0xbfbc8000.
> Allocation retried at 0xb4400000.
> Check SHMBASE is consistent with the value in
> $INFORMIXDIR/etc/onconfig.std.
> Consider using a different SHMBASE value in your ONCONFIG file.
> If shared memory segments are not allocated in increasing address order,
> memory block allocation performance may degrade.
>
> 13:43:05 Contiguous shared memory segment allocation failed at> 0x6c144000.
> Allocation retried at 0xb3c00000.
> Check SHMBASE is consistent with the value in
> $INFORMIXDIR/etc/onconfig.std.
> Consider using a different SHMBASE value in your ONCONFIG file.
> If shared memory segments are not allocated in increasing address order,
> memory block allocation performance may degrade.
>
> 13:43:05 Dynamically allocated new message shared memory segment (size
> 1616KB)
> 13:43:05 Contiguous shared memory segment allocation failed at> 0x6c144000.
> Allocation retried at 0xb3400000.
> Check SHMBASE is consistent with the value in
> $INFORMIXDIR/etc/onconfig.std.
> Consider using a different SHMBASE value in your ONCONFIG file.
> If shared memory segments are not allocated in increasing address order,
> memory block allocation performance may degrade.
>
> 13:43:05 Dynamically allocated new message shared memory segment (size
> 1616KB)
> 13:43:05 Contiguous shared memory segment allocation failed at> 0x6c144000.
> Allocation retried at 0xb2c00000.
> Check SHMBASE is consistent with the value in
> $INFORMIXDIR/etc/onconfig.std.
> Consider using a different SHMBASE value in your ONCONFIG file.
> If shared memory segments are not allocated in increasing address order,
> memory block allocation performance may degrade.
>
> 13:43:05 Dynamically allocated new message shared memory segment (size
> 1616KB)
> 13:43:06 IBM Informix Dynamic Server Version 7.31.UD8 Software Serial> Number AAC#J874023
> 13:43:06 IBM Informix Dynamic Server Initialized -- Shared Memory> Initialized.
> 13:43:06 Physical Recovery Started.
> 13:43:06 Physical Recovery Complete: 0 Pages Restored.
> 13:43:06 Logical Recovery Started.
> 13:43:09 Logical Recovery Complete.> 0 Committed, 0 Rolled Back, 0 Open, 0 Bad Locks
>
> 13:43:10 Dataskip is now OFF for all dbspaces
> 13:43:10 Checkpoint Completed: duration was 0 seconds.
> 13:43:10 Checkpoint loguniq 622486, logpos 0x380018
>
> 13:43:10 On-Line Mode
> 13:43:10 Affinitied VP 5 to phys proc 5
> 13:43:11 Affinitied VP 1 to phys proc 2
> 13:43:11 Affinitied VP 4 to phys proc 4
> 13:43:11 Affinitied VP 7 to phys proc 7
> 13:43:11 Affinitied VP 6 to phys proc 6
> 13:43:11 Affinitied VP 3 to phys proc 3
> 13:48:39 Checkpoint Completed: duration was 0 seconds.> 1
>
Neil Truby wrote:
> I shoul add that I've checked the SHMBASE and stune kernel parameters for
> the server and they all accord with the release notes.
>
> "Neil Truby" <neil.truby@ardenta.com> wrote in message
> news:2uaih3F28uitrU1@uni-berlin.de...
>
>>7.31 UD8 on SCO Unixware 7.1.1
>>
>>Our customer has just upgraded from UC6 -> UD8, and is seeing these shared
>>memory segment messages at start-up time. We've not seen them before on
>>UD8 on other OSs (AIX or Linux).
>>
>>has anyone else seen them, know what they mean, know if they are important
>>or know hopw to avoid them?
I dunno whether it is still valid, but it used (back in the early 90s)
to be necessary to ensure that the maximum shared memory segment size
was an exact multiple of 4 MB because the o/s would allocate shared
memory segments on 4 MB boundaries, and to achieve contiguous
addresses, therefore, the shared memory segments had to be (a multiple
of) 4 MB each. This applied to SCO Xenix and earlier versions of Unix
(and the mighty 386 or 486); I've no idea whether it applies to
Unixware running on Pentium class CPUs.
>>======================================================================
>>Wed Oct 27 13:42:50 2004
>>
>>13:42:50 Event alarms enabled. ALARMPROG =
>>'/usr/informix/etc/log_full.sh'
>>13:43:05 DR: DRAUTO is 0 (Off)
>>13:43:05 Contiguous shared memory segment allocation failed at>>0xbfbc8000.
>>Allocation retried at 0xb4400000.
>>Check SHMBASE is consistent with the value in
>>$INFORMIXDIR/etc/onconfig.std.
>>Consider using a different SHMBASE value in your ONCONFIG file.
>>If shared memory segments are not allocated in increasing address order,
>>memory block allocation performance may degrade.
>>
>>13:43:05 Contiguous shared memory segment allocation failed at>>0x6c144000.
>>Allocation retried at 0xb3c00000.
>>Check SHMBASE is consistent with the value in
>>$INFORMIXDIR/etc/onconfig.std.
>>Consider using a different SHMBASE value in your ONCONFIG file.
>>If shared memory segments are not allocated in increasing address order,
>>memory block allocation performance may degrade.
>>
>>[...snippage...the server appeared to come up...]
--
Jonathan Leffler #include <disclaimer.h>
Email: jleffler@earthlink.net, jleffler@us.ibm.com
Guardian of DBD::Informix v2003.04 -- http://dbi.perl.org/