Comments
11 comments
-
Hi,
As per my previous post we uninstalled Red-gate and the server did not shutdown unexpectedly last night, so the start of the backup job causes the server to shutdown unexpectedly.
What settings can I check or could it be as simple as a corrupt install?
Reinstalling today to see the outcome
Thanks -
Hi russellfoggy,
This isn't something that I have heard of before. If you run an Add-Hoc backup rather than a scheduled task does it complete or are there complications? Are there any error messages in the SQL Server log, the SQL Server Agent log or the Application/System log etc?
Thanks, -
Hi,
We reinstalled the client again this morning and ran a backup which caused the server to shutdown.
I then unflagged the 'threads' option and the backup ran through without crashing the server.
How could the thread option do that?
Thanks
Russ -
Hi,
We just checked the System and SQL log and there are no error messages prior to the shutdown.
Thanks -
The 'threads' option tells SQL Backup to back up to multiple devices but store the compressed data in a single file. Could you please run the following tests to help us isolate the problem:
- run a native backup to multiple devices e.g.BACKUP DATABASE [xx] TO DISK = 'c:\file01.bak', DISK = 'c:\file02.bak'
- run a SQL Backup backup to multiple devices and files e.g.EXEC master..sqlbackup '-sql "BACKUP DATABASE [xx] TO DISK = [c:\file01.sqb], DISK = [c:\file02.sqb] WITH INIT"'
- run a SQL Backup backup to multiple devices and a single file e.g.EXEC master..sqlbackup '-sql "BACKUP DATABASE [xx] TO DISK = [c:\file01.sqb] WITH THREADCOUNT = 2, INIT"'
Thanks. -
Hi Peter,
We ran those backups with no issues, pls see below:
Code:
BACKUP DATABASE [xx] TO DISK = 'c:\file01.bak', DISK = 'c:\file02.bak'
- run a SQL Backup backup to multiple devices and files e.g.
Ran successfully
Code:
EXEC master..sqlbackup '-sql "BACKUP DATABASE [xx] TO DISK = [c:\file01.sqb], DISK = [c:\file02.sqb] WITH INIT"'
- run a SQL Backup backup to multiple devices and a single file e.g.
Ran successfully
Code:
EXEC master..sqlbackup '-sql "BACKUP DATABASE [xx] TO DISK = [c:\file01.sqb] WITH THREADCOUNT = 2, INIT"'
Ran successfully -
I expected the last command to fail. Could you please post the backup command that's being used by the job?
Thanks. -
Hi Peter
Script below, I turned on the 'threads' before I copied the script for you because when that is selected we get the shutdown. With the threads turned off we have had 2 days with backups and the server has not shut down
This really is strange because I have been running Red-gate for years now on about 50 servers and never seen this
--Full script
EXECUTE master..sqlbackup N'-SQL "BACKUP DATABASES [*] TO DISK = ''C:\SQL BACKUPS\<AUTO>.sqb'' WITH COMPRESSION = 2, ERASEFILES = 5, MAILTO = ''Terry-Sue.vanRensburg@bidvestbank.co.za; Morne.vandermerwe@bidvestbank.co.za'', THREADCOUNT = 12, VERIFY"' -
Does it happen for a particular database, or for any database that you back up using 12 threads? If the former, could you please run the earlier tests on that specific database? If the latter, could you please determine what number of threads trigger the shutdown, then run the earlier tests using that same number of devices/threads?
Thanks. -
Hi Peter,
We can try to establish on what thread number we have the failure....only prob is that the server is remote and an operational branch server and when we have an unexpected shut down someone has to be on site to physically power off the server to get it up again.
Other than what you have suggested can we try something else?
I made mention in my original post that we were getting this on about 7 new HP ML350 servers, it looks like those have 'settled' down (without us changing anything) -
When you use multiple backup threads using the THREADCOUNT option, you will increase the memory requirements for both SQL Server (to read and store the backup data) and SQL Backup (to process the backup data coming from multiple threads).
I guess the next best bet would be to check your Windows memory settings e.g. PAE, 3GB switches, AWE settings etc.
Thanks.
Add comment
Please sign in to leave a comment.
Whilst rolling out new branch servers, HP ML350 Windows 2003 SP2 we have been deploying Red-gate backup 5.3 to backup the SQL DB to a local drive
We have been getting unexpected server shut downs which seem to occur @ the same time the Red-gate job is scheduled to run. The servers would do an unexpected shutdown and just hang until they are physically power cycled.
Last night as a test we changed a branch server with a job normally scheduled @ 11pm (also when the shutdown occurs) to 12am and that is when the shutdown occured as well...12am and not 11pm.
As a further test I have uninstalled Red-gate from the server to see what happens tonight.
Anyone have this issue before?
Thanks
Russell