Here's the modified reply:
We support Partial backups via scripting. Simply add the @read_write_filegroups parameter to the backup statement.
For example:
execmaster..xp_backup_database @database='foo', @filename='c:\foo.bkp', @read_write_filegroups=1
You could also have a look at Fast Compression as it will only back up database changes using differential backups, preventing the redundant backup of unchanged data.
Thanks.
--
David
Differential backups will backup only the changes which took place after the the full backup. We are trying to avoid taking full backup of the entire database(2 TB) in which almost 90 percent of data will never change and which is time taking process(almost 7 hrs). By using this partial backups feature we can reduce the time to 30 mins.
Thanks,
That's true. My point was that you could take that initial Full (since you'll need at least one backup of the read-only data for safe-keeping), then set a long Full Backup Internal (could even be 365 Days) in Fast Compression, and set a high Threshold for data change to ensure that Fast Compression continues to take differential backups for as long as possible. I agree though that Partial backups may be a better solution for your database, but Fast Compression can be more automated from LiteSpeed in terms of management. Either way you go, know we do support both options.
Thanks.
Thanks David. I tried to test the partial backups on a test server and I was not able to do. The following are the steps which I did -
I have a database with five filegroups and I made 3 of them as read only. I took the full backup ( two_full.bkp) of the database, then i created some tables and again took the partial backup(two_partial.bkp) of the database. Created one more table and took differential backup(two_diff.bkp). I dropped the database, and tried to restore it back.
Restore sequence-I restored the two_full.bkp with norecovery, then tried to restore two_partial.bkp but I was not able to do it as I cannot continue to the next screen....I am attaching the screen shot. I was not getting the option next to proceed to next screen or to enter the encryption password.
Thaks a lot!!! It works.
The console is not partial backup aware. We are having discussions on what we may add to the next release. In the meantime, you'll need to restore from script on that database. For example:
exec master..xp_backup_database @database='p', @filename='c:\olr\p_full.bkp'
exec master..xp_backup_database @database='p', @filename='c:\olr\p_partial.bkp', @read_write_filegroups=1
exec master..xp_backup_database @database='p', @filename='c:\olr\p_partial_diff.bkp', @read_write_filegroups=1, @with='differential'
drop database p
exec master..xp_restore_database @filename='c:\olr\p_full.bkp', @with='norecovery'
exec master..xp_restore_database @filename='c:\olr\p_partial.bkp', @with='norecovery'
exec master..xp_restore_database @filename='c:\olr\p_partial_diff.bkp', @with='recovery'
Thanks.
We are using litespeed for backup and restore process. We are planning to convert some of the file groups to read only on a 2 TB database. Does the compression on read only works similarly or we can compress them more than before.
I apologize for not answering this sooner. I had a closer look at the configuration processing and unfortunately what are trying to do is not supported. I tried editing the INI file directly because the configuration editor does not allow blank extensions, but the driver is explicitly failing to load the configuration when an extension is missing.
Is HP Data Protector saving the backups to disk? In other words, is it simply kicking off a native BACKUP command? I ask because most third-party SQL Server backup products use the VDI API to capture the backup stream directly.If that's the case with Data Protector, the LiteSpeed Engine could not be used to accomplish what you want.
What you can do is use LiteSpeed to back up to disk (from your own jobs) and then use HP Data Protector file system agent to sweep up those backups to Data Protector. Doing so would not require the Data Protector agent for SQL Server.
Thanks.
Read-only filegroups would not make any difference when it comes to compression.However, if you're using partial backups or Fast Compression, the amount of data change will be far reduced since much of the database is read-only.
Thanks.
Thanks for your reply David
Googling HP Dataprotector and VDI API makes me suspect that it's using VDI API, so I guess Litespeed won't help me out here.
Backing up to disk using Litespeed, and having HP Dataprotector backing the files adds complexity so we will have to look for a different solution for this.
I new to litepseed, trying to get automated restore for restoring db from prod to uat. Is there any way we could use withmove feature so that i can relocate files or use exisiting file location for that database in uat server like in regular full restore?
Any help is appreciated
Thanks
Yes. Both the Automated Restore Wizard and the API support placing the Data and Log files in a different folder than the default. Please review the Automated Restore Wizard in the LiteSpeed console. You can then preview the script to see which parameters are used to perform the task (hint: @datafilepath and @logfilepath).
Thanks.
I just setup an Active/Passive cluster using Windows Server 2008 and SQL Server 2008, and the LiteSpeed install seemed to go as expected until it tried to push the MSI to the second Node and gave the message 'Unable to configure Node: [node name]'. I clicked OK then Next and it just tried it again with the same warning and yet again looping (OK, Next, OK, Next, etc). I finally just clicked Cancel where it gave the message saying Canceled Instance Configuration and that's it. Is there anything it missed? The Quest software appears to have been installed on the second node, so I know that part worked as expected, but what configuratino did it miss? This is a test cluster so I might fail over just to see what it does, but given I'm testing with the hope of installing this on our Production cluster later this week I'd like to know for sure what this means or how to avoid it.
Any thoughts?
Also note our Failover cluster has three nodes, but I only have SQL installed on two. The one it failed on is the Passive SQL Node, and the software appears to ahve installed fine on the third node which isn't setup with SQL. I'm not sure if that matters, but FYI if it does.
Thanks
Hello Sam,
Please run the Instance Configuration Wizard from the passive node, which the error indicates it can't configure, and the install should complete without any further errors. If this fails to work please open a support case.
In most cases you should have no issues installing LItespeed to a cluster. However, If you have any problems with the Instance Configuration Wizard during installation – skip the wizard (press Cancel) and complete the installation.Then after installation completes run the Instance Configuration Wizard from the Start menu and complete it.
Hope this helps,
Thanks Ben,
I ran the configuration wizard on the second node and it reran the installation on the first Active node then gave the Unable to configure error on Node 3. This node doesn't have SQL installed, so can I ignore this? I'm installing a second passive node on it now just to see if this resolves the installation error, but what if anything would we loose if I cancelled installation at this window without making it to FINISH? Or what scenarios might cause it not to be able to configure the passive node though it can install the MSI? Just curious.
Thanks.
"This node doesn't have SQL installed, so can I ignore this?"
Yes, you can ignore this error in your case since the wizard is simply trying to install the "litespeed" DLLs into the SQL instance's install folder (binn directory). Which won't exist on the 3rd node in your installation. Please note the proper "dev recommend" way to resolve this issue is to run the Instance Configuration Wizard on this 3rd node too.
"what scenarios might cause it not to be able to configure the passive node though it can install the MSI? "
Cluster Installations normally fail due to permission issues with the account being used by the installer to run instance configuration and the version of the Windows operating system Litespeed is being installed to. These accounts can vary between the logged in user's account, NT system account , or the SQL account depending on what phase the installer is running (copying files, quering for SQL instances, updating stored procedures) However, to be certain what the problem is we would need to look at the installers log files.
Hope this helps,
Well this is odd, I installed a new SQL Cluster node (Passive) on the third node, then ran the configuraton wizard from the second node to see if it got picked up. Here's the result which is an error box without any error.
clicking OK gives me this:
Then clicking OK just returns back to the Summary page, and clicking Next just does this again, blank error then the Coinitialize error. If I just need to run this on each node then i'm fine with that, but is there any harm with it trying to rerun the configuration on other nodes that have already been configured?
Interesting, I'm not sure at what point this happened, but the cluster failed over to the third node after installing LiteSpeed. That was qutie unexpected since I didn't think installing litespeed would cause a faileover.
I'm giong through the logs now to see why, but anything in LiteSpeed that would cause a failover?
Hello Sam,
First I would recommend opening a support case for these installation issues as they are a little beyond what can be dealt with, via the forums, in a timely manner....however I can answer your questions...
is there any harm with it trying to rerun the configuration on other nodes that have already been configured?
No, you can rerun instance configuration at anytime. It will go thru the steps of configuring all nodes again, but this is not harmful.
Moving forward we are looking into the ability to check off nodes, where Litespeed is already installed, so they are not include in the re-running of instance configuration.
I'm giong through the logs now to see why, but anything in LiteSpeed that would cause a failover?
Not during an install. To confirm we would need to see the application and system event logs (in EVT format)of the node that failed over, Litespeed installer logs from the node where instance configuration was ran and, if possible, any SQL\windows dump files created on the active node which failed.
Just as an "FYI"...I have seen the error shown in your screenshot before, but I would need to confirm its the same issue first. So, again at this point, please open a support case.
thanks,
Thanks for th reply. I think the problem was probably due to me not having SQL Server installed on one of the cluster nodes when I installed LiteSpeed. I ended up installing SQL on the 3rd node as another passive node and the cluster configuration finished. I'm about to uninstall Litespeed and reinstall it now that i have SQL installed on all three nodes of the cluster. I'll report back, but maybe there could be something in future releases to verify SQL Server is instaleld before it tries to install LiteSpeed.
Also this is just a test cluster which we'll probably delete after this test is complete,so I'd rather not contact support for it, but after I uninstall and reinstall if I run into problems I might given I don't want to run into this when we install this in production. But again hopefully i think the problem was due to me not having SQL installed on the third node.
thanks again for all your help and confirmations ...