Machine Down. Replication question
Posted in 2003
Topics: Storage & Space Management, Logging & Checkpoints
We had to break replication a few weeks ago because of a problem in 9.2 UC5.
Yesterday I tried to bring my secondary up to do a sanity test. Heres what I
got in the log...
09:38:45 Informix Dynamic Server 2000 Initialized -- Shared MemoryInitialized.
09:38:45 Physical Recovery Started at Page(1:28060).
09:38:45 Physical Recovery Complete: 0 Pages Examined 0 Pages Restored.
09:38:45 Logical Recovery Started.
09:38:45 10 recovery worker threads will be started.
09:38:45 Logical Log File not found.
09:38:45 Cannot Rollforward from Checkpoint.
09:38:45 Informix Dynamic Server 2000 Stopped.
Here is the solution in the manual ...
Logical-log File not found
Cause: The checkpoint record in the root dbspace reserved page is corrupted.
Action: Perform a data restore from dbspace backup.
There is no root dbspace backup that I can restore from. If I backup from
the primary machine, can I restore that onto the secondary and have it work
? There is no critical data on that machine right now. All I want is for
Informix to come up so that I can try bringing replication back up. For that
I need to do a level 0 restore anyways.
Any ideas or thoughts ?
On Mon, 25 Aug 2003 09:58:54 -0400, Jack A wrote:
See below:
> We had to break replication a few weeks ago because of a problem in 9.2
> UC5. Yesterday I tried to bring my secondary up to do a sanity test.
> Heres what I got in the log...
>
> 09:38:45 Informix Dynamic Server 2000 Initialized -- Shared Memory> Initialized.
> 09:38:45 Physical Recovery Started at Page(1:28060). 09:38:45 Physical> Recovery Complete: 0 Pages Examined 0 Pages Restored. 09:38:45 Logical
> Recovery Started.
> 09:38:45 10 recovery worker threads will be started. 09:38:45 Logical> Log File not found. 09:38:45 Cannot Rollforward from Checkpoint.
> 09:38:45 Informix Dynamic Server 2000 Stopped.>
> Here is the solution in the manual ...
>
> Logical-log File not found
> Cause: The checkpoint record in the root dbspace reserved page is
> corrupted. Action: Perform a data restore from dbspace backup.
>
>
> There is no root dbspace backup that I can restore from. If I backup
> from the primary machine, can I restore that onto the secondary and have
> it work ? There is no critical data on that machine right now. All I
> want is for Informix to come up so that I can try bringing replication
> back up. For that I need to do a level 0 restore anyways.
That's EXACTLY what's required. The secondary was down so long that the
first logical log file that it missed is no longer online on the primary.
The only solution is to take a level 0 archive of the primary and
perform a physical restore on the secondary that will reestablish the
parity between the two servers as if you were setting up HDR for the
first time. Basically follow the manual as if it is a newly replicated
server.
> Any ideas or thoughts ?
Art S. Kagel