After Image apply problem

[Bill Burke]

We are using after-image extents to keep a warm-spare.

Long story short, every 20 minutes the current AI extent is closed and copied to the alternate server. The process works 99% of the time. The other 1% I get this:

********************************************
15:10:06 rfutil -C roll forward session begin for BUMTG on CON:. (451)
15:10:07 Begin Physical Redo Phase at 0 . (5326)
15:12:11 Physical Redo Phase Completed at blk 8908 off 2638 upd 113829. (7161)
15:12:11 Beginning roll forward of after-image file D:\Data\Replicat\Uni50Prd\unifipc_10-16-2001_150505-S15.ai. (660)
15:12:11 SYSTEM ERROR: Physical redo, BKUPDCTR=47, note updctr=49. (1028)
15:12:11 SYSTEM ERROR: Roll forward of file D:\Data\Replicat\Uni50Prd\unifipc_10-16-2001_150505-S15.ai did not complete normally (1046)
15:12:11 SYSTEM ERROR: End of last note read is at file offset 174155 (1048)
15:12:11 User 0 died with 1 buffers locked. (2523)
15:12:11 rfutil -C roll forward session end. (334)
*************************************************
That's the long way to say it didn't work.

The after image is the correct sequence. Recopying the ai file doesn't help. Restoring the last backup and all AI's doesn't help either.

This is Progress v9.0b on NT

Any ideas folks?

thanks
Bill

 

[Unregistered]

Dear Author


Have you tried the "Roll forward Retry" option after the error occured during roll forward.

Check it out this option.


I will revert to you because I had faced the same situtation sometime recently.


Contact me on rehan_ahmed@cdcpak.com


Thanks
Rehan Ahmed Khan
Senior DBA

 

[Bill Burke]

Hi Rehan,

Thanks for the idea of using the Retry option. I tried that today when I had an apply fail but it didn't work. Here are some additional related messages that I have found when it happened today:

PROGRESS PATCH Version 9.0B44 as of October 31, 2000
** The database was last changed Tue Oct 30 15:40:06 2001. (831)
** The after-image file expected Tue Oct 30 14:55:05 2001. (832)
** Those dates don't match, so you have the wrong copy of one of them. (833)


I have no idea why this is happening. It is the correct sequence. I don't know what else could be the problem. How am I a supposed to be able to maintain a warm spare with the garbage happening!!!!!!

Originally posted by Unregistered
Dear Author


Have you tried the "Roll forward Retry" option after the error occured during roll forward.

Check it out this option.


I will revert to you because I had faced the same situtation sometime recently.


Contact me on rehan_ahmed@cdcpak.com


Thanks
Rehan Ahmed Khan
Senior DBA

 

[Norman Biggar]

Bill,
Are you sure that the warm backup database is being kept in a "virginal" state?
If anyone tries to do something other than apply after image extents, problems can arise.
We had Doug Merrit from Progress visiting a few months back, and he advised:
1. Recreate your warm backup every time you perform a full backup.
2. The only processes to be performed on the warm backup should be after image roll forwards.

Looking at the database .lg file would let you see if someone has sneakily pro'd it or truncated its bi when you weren't looking.

 

[Bill Burke]

Hi Norm,

Thanks for your response!

I'm fairly certain that the warm spare is "virginal", as I am the only one who does anything with it. I gave some thought trying to remember if I had done anything to it yesterday when it happened, but fairly certain that I had not.

In fact, in times past when this occurred, I would attempt to fix it by restoring the backup and then immediately applying the AI extents only to have it fail at the same point.

Good thought though, and it may help me with another issue that I am having.

Thanks again,
Bill

Originally posted by Norman Biggar
Bill,
Are you sure that the warm backup database is being kept in a "virginal" state?
If anyone tries to do something other than apply after image extents, problems can arise.
We had Doug Merrit from Progress visiting a few months back, and he advised:
1. Recreate your warm backup every time you perform a full backup.
2. The only processes to be performed on the warm backup should be after image roll forwards.

Looking at the database .lg file would let you see if someone has sneakily pro'd it or truncated its bi when you weren't looking.

 

[wanwei]

Same issue on progress 9.1E with patch 4 on HP UNIX 11.11 64bit
Roll forward Retry Activated. (6804)
SYSTEM ERROR: Rollforward Apply AI note, BKUPDCTR=96642074, note updctr=96652056. (1028)
area 6, dbkey 2944, record type 37 (8999)
433 notes were processed. (1634)
0 in-flight transactions. (3785)
1 transactions were started
1 transactions were ended
At the end of the .ai file, 1 of them were still active
SYSTEM ERROR: Roll forward of file sinput.ai1 did not complete normally (1046)
SYSTEM ERROR: End of last note read is at file offset 25802 (1048)