RE: restoring possibly damaged logs
Posted in 1999
Topics: Backup & Restore, Storage & Space Management, Logging & Checkpoints
As far as the informix database is concerned, your data will remain consistent
and good through the rolling forward of all logs. You will not end up with
a corrupt chunk. Whether or not the operations done by your programs after
the
chunk went in error is a question which needs to be asked. This is a
determination of how well these programs handle failing SQL. You can have
inaccurate data in your database due to programs having problems handling
(or not handling) the errors caused by a bad chunk, but after the
restore the tables will not be corrupt from the point of view of the Informix
database.
Hope this helps
Will
>===== Original Message From tam.mclaughlin@net.ntl.com =====
>Rudy Fernandes wrote:
>
>> Yuri Dovgart wrote:
>>
>> > Tam McLaughlin <tam.mclaughlin@net.ntl.com> wrote in message
>> > news:385DFDD6.565979E4@net.ntl.com...
>> > > We had a situation where there was a power cut, UPS kicked in and shut
>> > > down informix.
>> > ...
>> >
>> > Generally speaking, you wrong. If your chunk is corrupted (for example,
one
>> > of it's pages), this means engine problem, not transaction. Anyway, it
>> > depends on what happends to your chunk. In your case you will lost all
>> > transactions that are in your logs on disk.
>>
>> My understanding is slightly different. ontape -c would have copied all
full
>> logical logs to tape. The log that was current at the time of the crash
would
>> be the only one missing from the tape. Consequently, all transactions that
>> were committed in that log would be lost (in addition, of course, to the
ones
>> that were still open at the time of the crash!).
>>
>> The correct procedure, of course, would have been to salvage that last log
>> (and others, to play safe!) onto a different tape for use during the
logical
>> restore.
>>
>> Rudy
>
>Sorry if I have not explained this properly. From the continuous log (ontape
-c)
>we were able to restore to the last log on tape before the crash. After the
crash,
>the server came back up and automatically started informix but not a
continuous
>log. When I came in on the morning, some cron jobs had been running during
this
>tme (2:30 am - 9 am). I shut down the engine and eventually done a restore.
The
>programs running after the crash would have generated transactions and filled
up a
>few logical log files despite their being a corrupt chunk. At the restore I
had
>the option to backup these current logs for use in the restore later but I
did not
>roll forward the current logs (which would have been on a second log log
tape) as
>I thought it best to restore to the log log file just before the
>crash. This way i def knew there would be no corrupt chunks. But would there
have
>still been a corrupt chunk or could I be 100% positive that the chunk was ok
and
>the integrity of the system was ok if i had restored the log logs completed
while
>there was a bad offline chunk ?
------------------------------------------------------------
This e-mail has been sent to you courtesy of OperaMail, as a free service from
Opera Software, makers of the award-winning Web Browser, Opera. Visit us at
http://www.opera.com/ or our portal at: http://www.myopera.com/ Your free e-mail
account is waiting at: http://www.operamail.com/
------------------------------------------------------------