Comments
Sort by recent activity
You can include the VERIFYINTERVAL = number in seconds parameter as it was introduced in 6.5.2.9 but this would further delay the time your backup takes. I would expect that the CHECKSUM happens as part of the backup not the verify stage. Red Gate people will confirm this.
If you have Page Verify set to CHECKSUM as a database option I would think the CHECKSUM in the backup could be redundant. True you would need to read every page in the database with Page Verify as CHECKSUM first to know that you had a good database before you run backups without CHECKSUM as all the newly writen pages are CHECKED because of the Page Verify setting.
We have a DB that is about 170Gb of data that we compress with a setting of 2 to 2 files. The backup takes 45 minutes and the verify messages come out 30 minutes later as they only appear at the end of that process.
If I remember correctly you try and compress this database with a level of 4 and haven't been able to complete a backup for at least a week unless that was fixed.
Chris / comments
You can include the VERIFYINTERVAL = number in seconds parameter as it was introduced in 6.5.2.9 but this would further delay the time your backup takes. I would expect that the CHECKSUM happens as...
I Just updated a server that had a C:\Program Files (x86)\Red Gate\SQL Backup folder rather than a SQL Backup 6 folder and it updated the SQL backup folder.
Chris / comments
I Just updated a server that had a C:\Program Files (x86)\Red Gate\SQL Backup folder rather than a SQL Backup 6 folder and it updated the SQL backup folder.
Chris
The product works as expected. I just didn't expect it to create a new folder and then not use it.
Chris / comments
The product works as expected. I just didn't expect it to create a new folder and then not use it.
Chris
Here's a tip if you didn't have the problems that you have. I saw this technique at PASS a few years ago.
They set up their new server with all the jobs etc of the old server. They made sure that it worked fine. They then took a full backup and restored it with norecovery. They then took differentials and restored the last with norecovery and finished it off by running a log backup and restoring with recovery. They were doing a bigger job than you by upgrading versions and moving to a cluster so that would have made your task easier even if you upgraded the SQL build. But you have no backups so that would not work.
What I determined from that was to have the system databases on a different drive and LUN. If we ever had to do that I would shutdown SQL on the old server, unpresent the drive from the old server, present it to the new server and attach the database files. If it was going from the same build of SQL then nothing more would be needed other than maybe point clients at the new server.
Might even be able to try this in the near future.
Chris / comments
Here's a tip if you didn't have the problems that you have. I saw this technique at PASS a few years ago.
They set up their new server with all the jobs etc of the old server. They made sure that i...
If you haven't tried the DBCC CHECKDB then it could be the CHECKSUM. It could also be the verify. If your DB is corrupt good luck. You will need Paul Randal's help http://sqlskills.com/blogs/paul/.
hopefully you have the last full backup at hand and if you could run a differential without checksum and verify and you had somewhere else to restore them to you could run the DBCC CHECKDB.
Chris / comments
If you haven't tried the DBCC CHECKDB then it could be the CHECKSUM. It could also be the verify. If your DB is corrupt good luck. You will need Paul Randal's help http://sqlskills.com/blogs/paul/....
I am sure that you have already tried this but.
Can you write anything to that drive?
Can you write anything to that drive using the Windows account that SQLBackup is using?
Can you do a log backup or a differential backup rather than a full?
Can you run DBCC Checkdb with PHYSICAL_ONLY on the database in question?
Chris / comments
I am sure that you have already tried this but.
Can you write anything to that drive?
Can you write anything to that drive using the Windows account that SQLBackup is using?
Can you do a log backup...
Do you use CLR routines and do you see any AppDomain 6 (AAA_BBB.dbo[runtime].5) is marked for unload due to memory pressure messages in the errorlog?
if you do then your problem could well be lack of contiguous memory in the vas.
Chris / comments
Do you use CLR routines and do you see any AppDomain 6 (AAA_BBB.dbo[runtime].5) is marked for unload due to memory pressure messages in the errorlog?
if you do then your problem could well be lack ...
There doesn't appear to be any fixes in the CU's raised after SP1 that talk about memory.
Red Gate SQLBackup uses the vas space in SQL which defaults to 256Mb in SQL2008R2. If this was changed you would see a line like this -g 400 in the SQL errorlog after lines like -l [image] \DBMS\MSSQL.1\MSSQL\DATA\mastlog.ldf to indicate where the master db log is after the last SQL startup.
I am not currently using SQLBackup on any SQL2008R2 servers but saw this error many times in SQL2005 running on W2K3R2 SP2.
A -g can be added to the startup of SQL but it would need to be stopped and restarted for it to apply.
Chris / comments
There doesn't appear to be any fixes in the CU's raised after SP1 that talk about memory.
Red Gate SQLBackup uses the vas space in SQL which defaults to 256Mb in SQL2008R2. If this was changed you ...
Unless you mean SQL2008R2 SP2 CTP I don't believe that version is freely available yet.
Is this 32 or 64 bit and how much memory does the server have and are you giving to SQL?
Chris / comments
Unless you mean SQL2008R2 SP2 CTP I don't believe that version is freely available yet.
Is this 32 or 64 bit and how much memory does the server have and are you giving to SQL?
Chris
Margaret,
You didn't mention the O/S or SQL version. This might help understand things a little better.
Chris / comments
Margaret,
You didn't mention the O/S or SQL version. This might help understand things a little better.
Chris