Video Screencast Help

BE 2012 Administration Course Question

Created: 19 Oct 2013 • Updated: 11 Nov 2013 | 5 comments
This issue has been solved. See solution.

Hello,

I am curently half way through the Backup Exec 2012 Adminitration Online Course Material and had a few questions that were not answered in the videos. If anyone could help out that would be great.

Question 1

Say there is a device pool in BE with 5 Disk Based locations and there are multiple locations online. If a backup job was to run, in what order does it pick what location to use? I believe it is alphabetical order…

Answer:

Yes, it will go in alphabetical Order

 

Question 2

DLM is run every 4 hours or when a Low space threshold is triggered on a disk. Does it run every 4 hours even if a Backup job is running to a disk that is online and being used by an active backup job?

Answer

DLM runs every 4 hours irrespective of whether any jobs are running, but apart from maybe a very minor performance hit will not affect the active backups even if it choosen to delete some files from the disk storage device you are currently using. Nite I think that triggering a DLM process by filling a disk does reset the 4 hour count for the regular background process.

 

Question 3

If an Entire Server is selected for backup with a C,D,E  Drive, and the E Drive was actually the backup destination, will Backup Exec automatically exclude the BEDATA Folder on the E Drive, or does it have to be manually excluded? I assume it will…

Answer:

As far as I can tell from a quick test this mornig we still backup the B2D data (although because we backup a snapshot, and force new bkf files for each set, it will not swallow its own tail by trying to backup the file that is actively being written to)

 

Question 4

If SDR was disabled on a server, and the server had to be recovered entirely from backup. How would this be done? I assume you would have to install the OS, and the restore the C and System State afterwards.

Answer:

Yes complete manual DR would be needed which starts by installing an oprating system with the same name (and preferably sam IP address) as a workgrouped machine (not a domain member), then install the BE componets (Full media server or RAWS depending on situation)

 

Question 5

Q: SDR basically requires all critical system components such as System State and C Drive. If I untick the server so that nothing is selected, right click the Simplified Disaster Recovery area and go "Select all necessary components for Disaster Recovery" it selects the C and System State, and unchecks everything else.

BUT

If I was to untick the server so nothing is selected, and MANUALLY tick the C and System State Selections, it does not Turn SDR On.

Why is that?

Answer: Will raise a ticket with tech support.

 

Question 6

Q: In the Material it says that you cannot perform a SDR of the BES.

Typically in my environment I run a SBS 2011 Server and a Terminal Server on the single ESXI Host.

I am using BE SBE with the BES on the SBS 2011 and a Agent on the terminal.

I thought that If I was backing up my SBS 2011 with SDR, and had a copy of the .dr file that I could simply purchase a new server, install ESXi, create a new VM, put in my SDRD, boot from that disk, and perform a restore from the Backup Drive (if it had a copy of the .DR file).

Is this not the case?

Answer:

Yes it is possible to SDR a BE Server, as long as its not running a dedup folder.

 

Question 7

Refer to the images below. Say I have a Server that has SDR turned on by default. The server has a C Drive and E Drive along with Exchange, SQL etc (It is a SBS 2011 Server).

On the E Drive there is a folder under E:\Shares called "_Archived Data and Backups [Not Backed Up]"

 

If I am to manually deselect this folder it turns SDR OFF.

Why does this happen? The data in the folder is not required for the system to operate so I wouldn’t think SDR would be disabled.

 

SDR Turn On.jpg

 

 

SDR Turned Off.jpg

 

Operating Systems:

Comments 5 CommentsJump to latest comment

Colin Weaver's picture

As I am not in the office I'll answer the ones I can without access to a Backup Exec Install etc

 

1) Yes as far as I am aware device in disk pool's are chosen alphabetically - although we are considering future options to make choices based on free-space.

2) DLM runs every 4 hours irrespective of whether any jobs are running, but apart from maybe a very minor performance hit will not affect the active backups even if it choosen to delete some files from the disk storage device you are currently using. Note I think that triggering a DLM process by filling a disk does reset the 4 hour count for the regular background process.

3) ...

4) Yes complete manual DR would be needed which starts by installing an operating system with the same name (and preferably sam IP address) as a workgrouped machine (not a domain member), then install the BE componets (Full media server or RAWS depending on situation)

5) ... (possiby a defect and as such might need a support case)

6) If BES = BE Server you can SDR a BE server - just not one running a dedup folder. If BES = something else then please elaborate. Your scenario implies that your SBS server is virtualized. The best solution to back this up and get a fast DR is to back it up using the VMware Agent (AVVI) using a system that is outside of the virtual environment as the media server. You will also need database agent license so that you can do GRT. However your SDR scenario should also work as long as dedup is not involved and as long as you have somehow got your backup storage device to be available (We do not support tape drive or USB disk access inside a VM as a backup target so I am wondering what your plan for a storage device is)

7) There is something on that volume that is consider to be critical to a server restore. It is not specifically the folder you are deselecting because as soon as part of a volume is defined as critical, then the whole volume is critical. Hence you cannot deselect the folder even if it is not the critical one. I am wondering (but am not sure) if because the configuration of shares are part of the system state, then do shares on a volume make a volume critical to SDR. We also do know that any service starting it's EXE files from a volume will make that volume be critical, so I hope you have not installed service based applications onto the same volume

 

Colin Weaver's picture

3) As far as I can tell from a quick test this mornig we still backup the B2D data (although because we backup a snapshot, and force new bkf files for each set, it will not swallow its own tail by trying to backup the file that is actively being written to)

 

Small clarification of point 1:

BE 2012 will choose first available device in pool but if more than one device is online it chooses by Alphabetical order within the online devices and 2012  will continue to use that device for all jobs until it is full (or offline)

BE 2010 and before has the capability to round robin choose within the online devices which should mean job1 goes to device1, job2 goes to device2, job3 goes to device3, job4 goes to device1 (assuming 3 online devices)

This is change in behaviour between 2010 R3 and 2012

Neither 2010 R3 or 2012 could make the choice based on disk space (which is the subject to a possible future enhancement.)

 

 

 

 

Peter Sheridan's picture

Hi Colin,

 

Thanks for your responses so far.

In regards to Question 6 - "(We do not support tape drive or USB disk access inside a VM as a backup target so I am wondering what your plan for a storage device is)"

What I actually do is pass through a dedicated USB2.0/3.0 controller directly to the VM itself. That way it acts the same way a normal server would. USB Drives are rotated daily and the SBS just sees it as a normal USB Drive (or more specifically a local disk).

I tried using the USB pass-through method through the hypervisor itself, but found that the transfer speeds were incredibly slow, and after talking to VMware they said it was never designed for large amounts of traffic like a backup would generate.

So I just stuck with passing through the PCI controller directly.

Unfortunately a lot of my clients sites are so small 1-30 Users, that it is hard to sell a dedicated Physical BE Server and the appropriate VRAY Edition Licence to backup the Virtual Host. I simply sell a Small Business Edition Licence that gives me rights to backup 3 Servers (Virtual or Physical) and do the pass-through method.

 

In regards to Question 7 - Is there a way to determine what BE sees as being critical for a SDR Restore?

Colin Weaver's picture

Hi Change of my answer to Question 5 as I did some digging

- When you select the whole machine it enables the automatic inclusion in the backup of any new resources that you might add to the same server at a later date. If you deselect the machine and then select individual components this feature is disabled.

- SSP/SDR status is linked to the above feature because we have no way to tell if in the future you are going to add a resource that also qualifies as a critical component. Hence you have to select the whole machine and then deslect the non-critical resources which leaves the above feature enabled and not choose to only select the critical resources.

 

Hence yes it is by design.

 

With Question 6 - whether or not you can get USB/Fiber/SAS or SCSI pass through to work with a storage device is up to you to test, we do not test or certify it hence not supported. We do know that tape pass through can initially appear to work but then all sorts of seemingly random device access issues start to affect backup reliability. If passthrough of the USB controller works for you then fine but you are using something we classify as an Alternate Configuration which we will only give limited support on unless you prove to us that you issue would occur with a completely physical setup. As such virtualized media servers are currently limited for official support to disk backup targets inside vmdk,disk backup targets on a NAS Share, disk backup targets on iSCSI presented into an initiator inside the VM and possibly iSCSI tape backup targets presented the same way. BTW the iSCSI disk target should be a physical device and not a software solution presenting iSCSI as we have seen issues with software iSCSI corrupting the backups sets (admittedly where a customer created a dedup folder in a software iSCSI target.)

 

On Question 7: Currently I am not aware of a way to determine why a volume is marked as critical for SSP/SDR - however both myself and at least one other colleague are trying to find out if there is something that can be easily understood by both support staff and customers/partners

SOLUTION
Peter Sheridan's picture

Hi Colin,

Thanks again for your detailed reponse. I appreciate it :)

If you ever hear of anything regarding question 7 please let me know. Thanks