IDS9.30HC5; HPUX11i running under an HP container.
No point discussing why we're on such an old IDS and telling me "you need to upgrade", this is just the way it is and I can't change it!
We have a down chunk on one of our testing instances; I think it occurred when I bounced the instance because of a hung process (we've had the odd strange unexplainable hang now we're permanently on the HP virtualised system that usually clears with an instance restart).
I'm pretty certain the chunk involved wasn't being accessed during the bounce and think the data in it should be sound.
I've tried marking it online using 'onspaces -O' but no dice. Obviously Informix tech support won't be interested in logging in to do whatever magic they do to mark it as online so, before we do an instance refresh from backups, does anyone know the magic keystrokes Informix used to use to mark a down chunk as up?
Ta
Hello Malc,
You are in trouble. A restore of the dbspace requires a rollforward and that
will most likely fail for the same reason. A pit restore of just before this transaction started will most likely succeed.
I would try this first on test though.
Otherwise you need to phone TS to have your chunk(s) and dbspace marked online
again.(that will leave an inconsistent transaction though!!!)
See you
Superboer.
Op maandag 8 september 2014 17:06:56 UTC+2 schreef Malc P:
> On Monday, 8 September 2014 14:09:02 UTC+1, wimpunk wrote:
>
> >
> > Isn't there anything in online.log telling you why it's not coming back?
>
> This is the online log at the time I restarted:
> 16:36:27 Informix Dynamic Server Initialized -- Shared Memory Initialized.
>
> 16:36:27 Physical Recovery Started at Page(34:202215).
> 16:36:27 Physical Recovery Complete: 159 Pages Examined 150 Pages Restored.
>
> 16:36:27 Logical Recovery Started.
> 16:36:27 10 recovery worker threads will be started.
> 16:36:27 Rollforward of log record failed. iserrno = 126
> 16:36:27 Log Record: log = 399804, pos = 0x5b536c, type = OLDRSAM:HDELETE(41),
> trans = 23
> 16:36:28 Ignoring the ONDBSPACEDOWN option during logical recovery. Dynamic Ser
> ver
> will NOT BLOCK at the next checkpoint
> 16:36:28 Assert Failed: Chunk 223 is being taken OFFLINE.
> 16:36:28 Informix Dynamic Server Version 9.30.HC5
> 16:36:28 Who: Session(11, informix@dev, 0, -2023062520)
> Thread(35, xchg_1.3, 8767cfbc, 1)
> File: rsmirror.c Line: 1870
> 16:36:28 Results: DBspace calc_dbs is disabled.
> 16:36:28 Action: Restore DBspace calc_dbs
:-p
--
Regards
Spokey
> On 8 Sep 2014, at 13:13, Jack Parker wrote:
>
>
> That will mark the down chunk as up?
>
> j.
>
>> On Sep 8, 2014, at 8:10 AM, Spokey Wheeler wrote:
>>
>> You need to upgrade.
>>
>> --
>> Regards
>> Spokey
>>
>>> On 8 Sep 2014, at 12:22, Malc P wrote:
>>>
>>> IDS9.30HC5; HPUX11i running under an HP container.
>>> No point discussing why we're on such an old IDS and telling me "you need to upgrade", this is just the way it is and I can't change it!
>>> We have a down chunk on one of our testing instances; I think it occurred when I bounced the instance because of a hung process (we've had the odd strange unexplainable hang now we're permanently on the HP virtualised system that usually clears with an instance restart).
>>> I'm pretty certain the chunk involved wasn't being accessed during the bounce and think the data in it should be sound.
>>> I've tried marking it online using 'onspaces -O' but no dice. Obviously Informix tech support won't be interested in logging in to do whatever magic they do to mark it as online so, before we do an instance refresh from backups, does anyone know the magic keystrokes Informix used to use to mark a down chunk as up?
>>> Ta
>>> _______________________________________________
>>> Informix-list mailing list
>>> Informix-list@iiug.org
>>> http://www.iiug.org/mailman/listinfo/informix-list
>> _______________________________________________
>> Informix-list mailing list
>> Informix-list@iiug.org
>> http://www.iiug.org/mailman/listinfo/informix-list
>
On Monday, 8 September 2014 14:09:02 UTC+1, wimpunk wrote:
>
> Isn't there anything in online.log telling you why it's not coming back?
This is the online log at the time I restarted:
16:36:27 Informix Dynamic Server Initialized -- Shared Memory Initialized.
16:36:27 Physical Recovery Started at Page(34:202215).
16:36:27 Physical Recovery Complete: 159 Pages Examined 150 Pages Restored.
16:36:27 Logical Recovery Started.
16:36:27 10 recovery worker threads will be started.
16:36:27 Rollforward of log record failed. iserrno = 126
16:36:27 Log Record: log = 399804, pos = 0x5b536c, type = OLDRSAM:HDELETE(41),
trans = 23
16:36:28 Ignoring the ONDBSPACEDOWN option during logical recovery. Dynamic Ser
ver
will NOT BLOCK at the next checkpoint
16:36:28 Assert Failed: Chunk 223 is being taken OFFLINE.
16:36:28 Informix Dynamic Server Version 9.30.HC5
16:36:28 Who: Session(11, informix@dev, 0, -2023062520)
Thread(35, xchg_1.3, 8767cfbc, 1)
File: rsmirror.c Line: 1870
16:36:28 Results: DBspace calc_dbs is disabled.
16:36:28 Action: Restore DBspace calc_dbs
On 09/08/14 13:22, Malc P wrote:
> IDS9.30HC5; HPUX11i running under an HP container.
> No point discussing why we're on such an old IDS and telling me "you need to upgrade", this is just the way it is and I can't change it!
> We have a down chunk on one of our testing instances; I think it occurred when I bounced the instance because of a hung process (we've had the odd strange unexplainable hang now we're permanently on the HP virtualised system that usually clears with an instance restart).
> I'm pretty certain the chunk involved wasn't being accessed during the bounce and think the data in it should be sound.
> I've tried marking it online using 'onspaces -O' but no dice. Obviously Informix tech support won't be interested in logging in to do whatever magic they do to mark it as online so, before we do an instance refresh from backups, does anyone know the magic keystrokes Informix used to use to mark a down chunk as up?
> Ta
>
Isn't there anything in online.log telling you why it's not coming back?
That will mark the down chunk as up?
j.
On Sep 8, 2014, at 8:10 AM, Spokey Wheeler wrote:
> You need to upgrade.
>
> --
> Regards
> Spokey
>
>> On 8 Sep 2014, at 12:22, Malc P wrote:
>>
>> IDS9.30HC5; HPUX11i running under an HP container.
>> No point discussing why we're on such an old IDS and telling me "you need to upgrade", this is just the way it is and I can't change it!
>> We have a down chunk on one of our testing instances; I think it occurred when I bounced the instance because of a hung process (we've had the odd strange unexplainable hang now we're permanently on the HP virtualised system that usually clears with an instance restart).
>> I'm pretty certain the chunk involved wasn't being accessed during the bounce and think the data in it should be sound.
>> I've tried marking it online using 'onspaces -O' but no dice. Obviously Informix tech support won't be interested in logging in to do whatever magic they do to mark it as online so, before we do an instance refresh from backups, does anyone know the magic keystrokes Informix used to use to mark a down chunk as up?
>> Ta
>> _______________________________________________
>> Informix-list mailing list
>> Informix-list@iiug.org
>> http://www.iiug.org/mailman/listinfo/informix-list
> _______________________________________________
> Informix-list mailing list
> Informix-list@iiug.org
> http://www.iiug.org/mailman/listinfo/informix-list
You need to upgrade.
--
Regards
Spokey
> On 8 Sep 2014, at 12:22, Malc P wrote:
>
> IDS9.30HC5; HPUX11i running under an HP container.
> No point discussing why we're on such an old IDS and telling me "you need to upgrade", this is just the way it is and I can't change it!
> We have a down chunk on one of our testing instances; I think it occurred when I bounced the instance because of a hung process (we've had the odd strange unexplainable hang now we're permanently on the HP virtualised system that usually clears with an instance restart).
> I'm pretty certain the chunk involved wasn't being accessed during the bounce and think the data in it should be sound.
> I've tried marking it online using 'onspaces -O' but no dice. Obviously Informix tech support won't be interested in logging in to do whatever magic they do to mark it as online so, before we do an instance refresh from backups, does anyone know the magic keystrokes Informix used to use to mark a down chunk as up?
> Ta
> _______________________________________________
> Informix-list mailing list
> Informix-list@iiug.org
> http://www.iiug.org/mailman/listinfo/informix-list