T.R | Title | User | Personal Name | Date | Lines |
---|
5051.1 | | M5::LWILCOX | Chocolate in January!! | Thu Feb 20 1997 09:32 | 9 |
| Yuck, this area looks really sick. I also saw note 1788 and the customer now
has a call into his fs rep. He is using MTI drives.
I am still interested in knowing if anyone has seen an instance of a checksum
error NOT being logged in the CPT.
Thanks.
Liz
|
5051.2 | | HOTRDB::LASTOVICA | Is it possible to be totally partial? | Thu Feb 20 1997 09:43 | 4 |
| since you indicated that the checksum error occured during the backup,
can I assume that he explicitly specified /CHECKSUM (it isn't the
default). Perhaps RMU/BACKUP doesn't update the CPT upon checksum
errors. It'd be easy for you to test.
|
5051.3 | please see bug 396737 | AVMSV1::EKREISLE | Erich Kreisler | Fri Feb 21 1997 03:49 | 7 |
| I had the problem that pages with an incorrect page number were not inserted
into the CPT.
See also note 4438.
Ciao,
erich
|
5051.4 | backup doesn't log it! | M5::LWILCOX | Chocolate in January!! | Fri Feb 21 1997 09:16 | 14 |
| <<< Note 5051.2 by HOTRDB::LASTOVICA "Is it possible to be totally partial?" >>>
>> default). Perhaps RMU/BACKUP doesn't update the CPT upon checksum
>> errors. It'd be easy for you to test.
Hmmmmm, I tried backup with and without /CHECKSUM after I purposely corrupted
a checksum via alter. It did NOT log the error in the CPT in either case!
I was surprised that it didn't. It sure does return the checksum error.
Is RMU/BACKUP suppossed to log the page in the CPT?
Thanks,
Liz
|
5051.5 | that's about what I expected | HOTRDB::LASTOVICA | Is it possible to be totally partial? | Fri Feb 21 1997 09:25 | 11 |
| >It did NOT log the error in the CPT in either case
With /NOCHECKSUM (the default), database page checksums are not
calculated or compared, so it is no wonder that this didn't even notice
and bad checksums. When you use /CHECKSUM, I believe that RMU signals
the error and stops the backup.
You've got the 6.1 listings on line (I believe that RMU is included), I
imagine that you can look at the code to determine if it ever intended
to update the CPT; My guess is that it doesn't even try. If you think
that this is a problem, a suggestion BUG might be in order.
|