How to best backup SharePoint 2013 with 2008R2 SQL and VMware?

[nackros]

Hi,

My customer´s environment is this:

- One server(2012) with SharePoint 2013 server(only one server in the farm)

- One server(2012) with MS SQL 2008R2(all sharepoint databases are here. this server also handles lots of other databases)

All servers are virtual using VMware.

The Backup Exec 2014 server(2008R2) is a physical standalone machine.

I need to backup everything, preferable using VMware Agent with GRT. The plan in to run Full backups of everything once a month, and incrementals daily.

One of the problem I have is that almost every Sharepoint Farm backups fail with the following errer: "Final error: 0xe0000363 - The Backup Exec SQL Agent was not used to create the last full, differential, or log backup of this database. You must use the SQL Agent to run a full backup before you can run a differential backup or transaction log backup."

Ofcourse I only backup with BE, nothin else. I interpret the message above as I do a VMware GRT backup of the database server first, the following backup using Sharepoint agent complains.

I would like to know what the best practises is to backup with a scenario like this?

I´m getting confused and suspect that I backup both the sharepoint anvironment and the sql environment twice which is stupid.

Some questions I have in my head are: Should I backup the SQL server using VMware Agent with SQL GRT? Since I want to restore individual databases other that Sharepoint. Or should I backup without GRT and do a backup using the SQL agent and select only the non-sharepoint databases?

And Should I do the same for the SharePoint server or not? Or what about backup the "farm" since that seems to backup databases too?

Please help me how I should set this up the best way.

Regards,

//Andreas..

[pkh]

Since your SharePoint databases are on a separate machine you cannot do VM GRT backups of your SharePoint farm. You need to run a separate job to backup the farm using the SharePoint agent.

[nackros]

Thanks for your reply.

So you mean that when I do the vmware backup of the sharepoint server I´ll just do that without GRT?

And backup the Farm separately.

How should I backup the SQL server then?

[pkh]

Yes to your first two questions. For your SQL VM, you would need to do GRT backups because you said you have non-SharePoint databases on it

[nackros]

Thanks!

But if I backup the SQL using GRT I get the error message in my first post when the Sharepoint farm backup runs.

How should I prevent that?

[pkh]

1. Do not backup SQL using other backup software or SQL maintenance. 2. If you are backing up the SQL VM and a job using the SQL agent, then do you VM backup before the full SQL backup, not between the full backup and any log/differential backup

[nackros]

1. I don´t use any other backup software, and the Maintenance jobs don´t run backups.

2. Now I´m backup up the SQL server using VMware GRT with SQL GRT enabled. And I´m using the exact same schedule when full/inc is done as the SharePoint backup.

Please explain a bit more. I need someone to tell me how I should set this up so everything runs without error messages. I want to run full backups once a month and incrementals every day in between. Should I change the backup time a bit for one of these jobs so that I now which starts first? In that case, which server?

[pkh]

See these two documents

http://www.symantec.com/docs/TECH137587

http://www.symantec.com/docs/TECH58674

Your schedule must be something like this

VM backup

SQL Full

SQL Log/Differential

It cannot be something like this

SQL Full

VM backup

SQL Log/Differential

[nackros]

Ok,

This is exactly how it´s configured now:

Server A(MS SQL with all sharepoint databases plus lots of other databases):

- Is backed up using VMware agent with GRT enabled for files and SQL, plus enabled log backups

- Full backup runs one saturday every month at 00:10

- Incrementals run every workday at 20:00

Server B(SharePoint 2013):

- Is backed up using VMware agent with GRT enabled for files only

- Full backup runs one saturday every month at 00:10

- Incrementals run every workday at 20:00

SharePoint 2013 Farm:

- Full backup runs one saturday every month at 00:10

- Incrementals run every workday at 20:00

What should I change/add to make this work? The error message is for the SharePoint Farm job only. all other backups are Successful.

[pkh]

ServerA is out of the picture because it does not have any Sharepoint resources.

However, for all 3 servers, your schedule is very bad because when you restore incremental backups, you need to restore the last full backup plus ALL the incremental backups since then.  Imagine if you need to restore something on the 30th of the month, your restore will take forever.

You should do full backups more frequently like weekly.

For the Sharepoint farm, assuming that you do a full backup on Sunday and incremental backups on Mon-Fri.  You can only run VM backups after the last incremental backup on Friday and the Sun full backup.  If you run VM backups during any other time, you will get the error message.

[nackros]

ServerA is out of the picture because it does not have any Sharepoint resources.

The SQL server holds all Sharepoint databases, so I would definately say it contains SharePoint resources, but it depends on what you mean with Sharepoint resources.

However, for all 3 servers, your schedule is very bad

The schedule works satisfactory for this customer. We seldom do restores, and everything is located on fast fibre SAN so a restore is very quick. We have done lots of test on this and 30 incrementals in a row is not a problem. Focus for this customer is the archivals for several years so everything is duplicated multiple times to other sites. So in this case the schedule is not very bad.

But with regards to the SharePoint farm, I don´t think I understand exactly what you mean. Do you mean that I should leave the Sharepoint farm and SQL server backups as they are now and only adjust the schedule for the Sharepoint server? Or have I misunderstand you?(this is probably the case ;))

[pkh]

If ServerA is holding the Sharepoint data, then you can do VM backup between the last Sharepoint incremental backup and the Full Sharepoint backup.  If you do VM backup at any other time, you will get the error message.

You have to adjust your VM backups so that it does not come between your full Sharepoint backup and your incremental Sharepoint backup.  If you are going to do full Sharepoint backups once a month, then you can only do your VM backups once a month.  This applies to both ServerA and ServerB.

[nackros]

Ok, thanks. I think I understand the problem now.

But since it´s not ok to only backup the server once a month I need to workaround this.

Would it work if I run the backup for Server A as it is now but skips GRT completely? And set up a new job that only backup the non-sharepoint databases using the SQL agent? that way it would only be the Sharepoint farm that backups the sharepoint databases. Am i correct?

[pkh]

No. It will not work. It is not GRT that is the problem. It is the VM backup

[nackros]

Ok, but since it´s not ok for me to only do vmbackups once a month, how should I solve this?

I think I´m back to my original question now. What´s the recemmended best practise when backing up SharePoint 2013 and SQL in a setup as mine?

[pkh]

Increase the number of times that you do full Sharepoint  or SQL backup.  Each time you do a VM backup you have to do a full Sharepoint or SQL backup.

[nackros]

Well I can´t do that. Full backups on everything once a month, and incrementals 5 times a week is the requirements.

I can´t believe that BE can´t handle this error. The backups itself seems to work for the most part, so this error message about "BE was not used for the last fullbackup.." is just a crappy bug. I don´t care if a different agent was used. It´s still BE2014 that does all backups. Can I suppress this error message somehow?

This error was present in BE2012 aswell and I hoped it would be fixed in 2014.. 2014 fixed lots of other 2012 issues, but not this.

[pkh]

I don't think this will be fixed because it is a SQL thing, not BE

[nackros]

Not the part when this error message plops up ofcourse. So I should be able to suppress this error message somehow? Or is that not possible? All backups and restores work. and since your solution is worse than mine this might be a workaround.

[pkh]

No. The error message cannot be suppressed