Quantcast
Channel: Software Communities: Message List - LiteSpeed for SQL Server
Viewing all 150 articles
Browse latest View live
↧

Re: Just installed LiteSpeed 7.1 on Active/Passive cluster, but Node 2 has error ..

May 21, 2013, 6:56 am
$
0
0

No problem I am happy to help.

 

Just as another "FYI" the error seen in your screenshot is most likely a bug we have identified as occuring when a cluster node is offline when instance configuration is run. This issue has been given the CR numbers 107658 & 105341 and will be addressed in the next version of Litespeed. This error can cause the instance config wizard to loop.... CR107571 addess that issue which is also fixed in the next version of Litespeed.

 

Hope this helps and please don't hesitate to contact us, via the forums or support, if you should have any issues or concerns with Litespeed.

↧
Search

How log is the LiteSpeed evaluation period for?

May 22, 2013, 12:12 pm
$
0
0

Silly question, but if we install LiteSpeed in test with the Evaluation license (entering no key) now long is it good for?  For the life of me I can't find this documented anywhere.

 

Thanks.

↧
↧

Re: How log is the LiteSpeed evaluation period for?

May 22, 2013, 12:23 pm
$
0
0

15 Days. If you need more time, please contact your sale rep for a trial extension (and a demo if you're interested).

 

Thanks.

 

--

David

↧

Re: How log is the LiteSpeed evaluation period for?

May 22, 2013, 12:34 pm
$
0
0

Sam,

 

You can also run this query to verify the Expire Date for the trial period : EXEC master.dbo.xp_sqllitespeed_licenseinfo

↧

Backing-up to UNC is failing, suggestions?

May 28, 2013, 7:38 am
$
0
0

I've scheduled a database backup through Litespeed Console to backup to a UNC path, but it's not working.  Here's the SQl from the step:

 

declare @script nvarchar(4000)
set @script = N'
exec master.dbo.xp_backup_database
@database = N''LiteSpeedLocal'',
@backupname = N''LiteSpeedLocal - Full Database Backup'',
@desc = N''Full Backup of LiteSpeedLocal on 5/26/2013 10:17:44 PM'',
@compressionlevel = 2,
@filename = N''\\cloudpath\mssql\LiteSpeedLocal.bak'',
@init = 0,
@with = N''STATS = 10'''

exec (@script)

 

and here's the error from the job:

 

Executed as user: Dev\sqlagentservice.

LiteSpeed(R) for SQL Server Version 7.1.0.1720 (64bit) Copyright 2013 Quest Software, Inc.

[SQLSTATE 01000] (Message 1)  Directory '\\cloudpath' could not be created [SQLSTATE 42000] (Error 61504).  The step failed.

 

Error SQLSTATE 42000 is Synax error or access violation, but when I log into a system as Dev\sqlagentservice I can view \\cloudpath\mssql\ and create/delete content.  Also I can backup to a local drive fine so I'm pretty sure Litespeed is working as it should. 

 

Any thoughts?  Thanks.

↧
↧

Re: Backing-up to UNC is failing, suggestions?

May 28, 2013, 8:06 am
$
0
0

Check your quotes. They should be single quotes, not double. If that does not help, then make sure the SQL Server service account has access to the destination folder.

 

Let us know what you find.

 

Thanks.

 

--

David

↧

Re: Backing-up to UNC is failing, suggestions?

May 28, 2013, 8:08 am
$
0
0

Sam,

 

You may wish to change the UNC path.  When I ran your script on my test server, adding the complete UNC path works:

 

declare @script nvarchar(4000)

set @script = N'

exec master.dbo.xp_backup_database

@database = N'LiteSpeedLocal',

@backupname = N'LiteSpeedLocal - Full Database Backup',

@desc = N'Full Backup of LiteSpeedLocal on 5/28/2013 7:59:45 AM',

@compressionlevel = 2,

@filename = N'\\Server_name\c$\Backup\2005\LiteSpeedLocal_201305280759.bak',

@init = 0,

@with = N'STATS = 10''''

exec (@script)

 

The SQL Server service account must have read/write/create permissions to the destination folder.

↧

LiteSpeed Console not connecting to instances -- suggestions?

May 29, 2013, 2:03 pm
$
0
0

We have a three node SQL 2008 cluster where Node 1 is Active (4 Instances), Node 2 is Active (4 Instances), and Node 3 is Passive for all 8 Instances.  We installed LiteSpeed 7.1 last week and configured each Instance where all 8 instances were showing-up in LiteSpeed Console.  Now today the 4 Instances on Node 2 aren't showing up.  Or rather it just says "Operation completed with errors: Object reference not set to an instance of an object".  I get this on LiteSpeed Console on both my system and on the NOde 2 server.  Though on the Node 2 server it says this "The local repository is not available.  This server instance does not have a local repository configured".  It was configured last week and working yesterday, so what would cause it to loose the configuration?  We have had no failovers and the LiteSpeedLocal database is on all Instances.  Also the Central Repository is on one of the Instances on Node 2 which isn't working, yet it works.

 

Any suggestions?  Thanks.

 

Sam

↧
Search

Compression of BLOBs (FileTables)

June 7, 2013, 2:36 pm
$
0
0

Hello everyone,

 

I think I know the answer but I thought I would ask the experts before going back to the customer.

 

Will SQL LiteSpeed compress a file table which stores JPEG's?  They are currently at 1TB DB size and he is very interested in speeding up the backup and compressing it at the same time.

 

Thank you,

Jamey

↧
↧

Re: Compression of BLOBs (FileTables)

June 9, 2013, 4:01 pm
$
0
0

JPEG files are already highly compressed. And once something is compressed, it's unlikely you'll get additional compression. Depending on how much other data exists in the database, you might still see good overall compression, but if a majority of the data is in the JPEG files, then LiteSpeed data compression may not help much. What we do recommend is customers  with binary data use Fast Compression at Compression level 1 (or higher if there is more non-JPEG compressible data in the database). Fast Compress reduces backups through its smart differential backup process and doesn't care if the data is compressible or not. You're likely to see an additional 70-85% reduction in backup footprint simply by moving to Fast Compression over nightly, full backups, and you won't have to back up all the JPEG files every night.

 

Thanks.

 

--

David Gugick

Product Management

Dell | Data Protection

↧

Re: Report from LiteSpeed on job durations?

July 1, 2013, 11:49 am
$
0
0

Yeah I'm figuring that out.  It appears the UI reports status from the jobstatus table but that was reported back from msdb and only gives the job name, status, and last run - no duration.  The only way to get duration is from the msdb sysjobhistory but because each job has multiple steps it's really confusing to try to aggregate the info together and get a real duration number for each job (and I profess I am not a DBA so just don't have the experience to write TSQL to mine the data - that's why I use LiteSpeed, Spotlight, and Foglight Performance Analysis to get the information I need to administer my platform).  Oh well - guess I'll be using the UI to get it together.  It would be nice if in the future LiteSpeed had a reporting feature that could report stuff like this (ie. job duration for a specific job across multiple instances\servers, or durations for all jobs in selected instances, or all jobs deployed and current status like enabled\disabled, and so on) - the Excel export is less than helpful unless it was re-written to export in a layout that's easier to work with.  If some feature was made available like that then I would still keep Excel but maybe offer a Crystal or PDF or Word output - or even use SSRS to generate the reports.  Thanks anyways for the help - your support is always great (even if I don't get an answer I like).

↧

LiteSpeed, SQL 2012 AlwaysOn, and Spotlight

July 2, 2013, 11:42 pm
$
0
0

I have a shiny new SQL 2012 AlwaysOn availability group, and am encountering some obstacles with the Litespeed implementation.  I could sure use some pointers, or maybe you could point me to a best practices document.

 

My configuration:

 

  • Two node AlwaysOn, with both nodes configured to be readable if they are secondary
  • Litespeed is running build 7.1.0.1720. 
  • Spotlight on SQL Server is version 9.5.0.6522
  • I use FastCompression backups, because I enjoy the logic that you guys employ to do differentials as warranted.
  • Maintenance plans were created inside LiteSpeed, and contain only the backup task.
  • I have FastCompression jobs configured for "All Databases" to run daily on both the Primary and Secondary replica.  The "AlwaysOn" setting is set to "Only back up the primary" (yuck)
  • Transaction Log maintenance plans run on "all databases" every 15 minutes.  The "AlwaysOn" setting is "Back up preferred"

 

A couple of bugs...

 

  • Fast Compression appears to be missing some AlwaysOn options.  Right now, the only options are "Only back up the primary", and "Back up none".  Neither of those are optimal, since you really want your backups coming from a ReadOnly secondary.  If I create a "Back Up Database" task, I have an option for "Back up preferred".... Is there a way to enable that option in Fast Compression?
  • Spotlight is throwing inaccurate alarms.  It is reporting that some databases on the AG have not had any differential backups in over a day, and others have not had log backups because they have never had a full backup.
    • This alarm appears to be incorrect.  If I inspect the backup history in the Litespeed backup browser, I can see that there are recent differential and full backups of these databases taken from the Primary replica, and log file backups taken from the secondary replica.



Also, a question on where I should be storing backups... My server names are SQL1, and SQL2.  I have backups for each server going to different folders (with "create a subdirectory for each database" checked).  The backup paths are \\FileStore\SQLBackups\SQL1 and \\FileStore\SQLBackups\SQL2.  The problem with this is that the backups for database that are members of the AG are scattered between two locations.  For example, for databases that are member of the AG, the FastCompression backups are in a subfolder of SQL1, and the transaction log backups (coming from the secondary replica) are in a subfolder of SQL2.  Won't this make a mess of trying to do a restore?

 

I could point all maintenance plans to a single folder (\\FileStore\SQLBackups), and then it would consolidate my AG database backups from both replicas into a single folder.  But then, what about databases that are not apart of the AG, but appear in each replica with the same name? (examples: msdb, master, model, LiteSpeedLocal, etc)

 

I looked in all of the obvious places for some kind of "Implementing LiteSpeed with AlwaysOn" document, but I haven't been able to find anything.

 

Thanks, -Rob

↧

Re: LiteSpeed, SQL 2012 AlwaysOn, and Spotlight

July 3, 2013, 6:01 am
$
0
0

Rob,

 

All good questions I'll try and answer below:

 

1. AlwaysOn backups on replicas: only Full and Transaction Log backups are supported on secondary replicas. Because of this SQL Server limitation you end up with limited locations for Fast Compression since it uses differential backups. You can either run them on the primary or skip them in the case where you want to implement custom jobs for those databases. It wold be possible to run the full backups on the secondary and the differentials on the primary, but Fast Compression does not support that (yet). As you noticed, when running Full backups, you get additional options.

 

2. You didn't ask, but I thought I'd mention it. Licensing SQL Server with AlwaysOn. My understanding is Microsoft grants a single free license for a secondary replica used in an AOAG, but it's only free if you only use it for failover - meaning you don't use it for backup or read-only querying. If it's active for those tasks, it may require a license.

 

3. Backup locations: Microsoft recommends you use a common share just for the convenience of having all backups in the same place, but it's not required. You're already doing this as best as you can given the db overlap. When you restore, and depending on which instance in the AG is being restored to, you will either select the option to restore from the database which pulls backup info from msdb, or restore from a disk device and select them manually. You can always use the disk device option if you want to manually select though. In LiteSpeed, currently, you would have to restore the full without recovery and then run the restore wizard again and restore all the Tlogs in one pass with recovery. In a future release you'll be able to restore all backups at once even if they were created on different replicas in the AG.

 

4. Spotlight: I'll pass your comments onto the Spotlight team.

 

 

Thanks.

 

 

--

David Gugick

Product Management

Dell Software | Data Protection

↧
↧

Re: LiteSpeed Console not connecting to instances -- suggestions?

July 3, 2013, 7:51 am
$
0
0

These types of cluster installation issues are best handled through our Support organization. They can provide much better service and tracking as there will be a need to get more detailed information about your environment and logs which you would not wanted posted publicly.

 

Thanks.

↧

Re: IO issues when restoring the database with listspeed

July 3, 2013, 8:02 am
$
0
0

Looks like this post was accidentally missed. I realize a lot of time has passed and I'm hoping you found a solution.

 

From what I can see, the issue is disk related. Is there a possibility of moving the backup to another drive? If not, can you restore to a different drive? SQL Server handles the restore process. LiteSpeed simply provides SQL Server the data for restore. If this is CPU related (although that's not usually the case on restore), you might try using a lower compression level for the backup. While Restore CPU used is far less than the CPU needed for a backup, there is a relationship and a compression level 8 backup will use more CPU on restore than a compression level 1 backup. On the other hand, if part of the issue is the Read IO on the backup file, then using a higher compression level for the backup may help. A compression level 8 backup, as an example, may be half the size of a compression level 1 backup -meaning, only 50% of the read IO is required on the backup file for restore.

 

Let us know if you came up with a solution.

↧
Search

Re: LiteSpeed, SQL 2012 AlwaysOn, and Spotlight

July 3, 2013, 12:56 pm
$
0
0

David,

 

Thanks for the thorough reply! 

 

It sounds like there is still some work to do in optimizing LiteSpeed for AlwaysOn.

  • Fast Compression options (full backups should come from secondary)
  • Managing backup file locations (and restoring from those files)

 

It sounds like you guys are fully aware of these issues and I am confident that you are working on them.  Can we hope to see them in 7.2?  If so, does your crystal ball have a ballpark time frame?

 

Regarding the licensing issue... Thank you very much for bringing that up.  I can't believe I made it through the entire purchase process, and I didn't know that.  Now I have a ton more AlwaysOn questions that are not necessarily related to LiteSpeed:

 

What is the best way to allocate my 8 core licenses?

  • Put all 8 cores on the primary, and make the secondary passive
  • Put 4 on the primary and 4 on the secondary, and implement Read-Only Routing
  • Use some other core mix with Read-Only Routing (eg: 6 on the primary and 2 on the secondary)

 

I have scoured the interwebs for some information on a performance comparison between these options, and have come up completely dry.  What's your gut thoughts on this?

 

Marking your original answer as "correct", but hoping you'll take a moment to reply anyway.  Cheers!

↧

Re: LiteSpeed, SQL 2012 AlwaysOn, and Spotlight

July 3, 2013, 1:27 pm
$
0
0

I would expect to see some improvements starting in 7.5 in August. In that release, you'll be able to grab all backups from disk device in the Restore Wizard in a single pass, regardless of which node in the AG ran the backup, and restore.

 

In the 8.0 release, we are planning to add support to be able to access all nodes in the AG during restore so we can dynamically pull back backup data from msdb (which is not shared between nodes in the AG). That will allow the LiteSpeed console to intelligently show you backups without you having to select them from disk.

 

I do not yet have a release date on whether we are going to add functionality to perform full backups on replicas and differential backups on the primary. Consider that feature TBD, but if it does show up it could be in 8.0 or 8.1.

 

Thanks.

 

--

David

↧
↧

Re: LiteSpeed, SQL 2012 AlwaysOn, and Spotlight

July 3, 2013, 1:33 pm
$
0
0

David, thanks for the peek ahead... I can't wait!

 

Have a great independence Day!

↧

Re: LiteSpeed, SQL 2012 AlwaysOn, and Spotlight

July 3, 2013, 1:48 pm
$
0
0

Regarding your answer #3:

" In LiteSpeed, currently, you would have to restore the full without recovery and then run the restore wizard again and restore all the Tlogs in one pass with recovery. In a future release you'll be able to restore all backups at once even if they were created on different replicas in the AG."

 

90% of the time, I am doing object level recovery.   What are the steps to do an object level recovery, with the files split like that?

↧

Re: Litespeed installation on cluster node

July 4, 2013, 9:04 am
$
0
0

LiteSpeed for SQL Server does not require a reboot on install, upgrade, or uninstall.

 

Thanks.

 

--

David Gugick

Product Management

Dell Software | Data Protection

↧
Viewing all 150 articles
Browse latest View live
Search