Incomplete Backups

[EHorwitz]

I'm using BESR 8.5 trial version on my desktop, backing up to a USB drive attached to a server.  For whatever reason, I've come in twice in the morning and found the backup has hung. Status says it may be as much as 95% complete. There are other issues on my desktop, so I have to reboot.

Since this is the trial version, I have no patches and don't know what the existing patches and updates address.

When I look at my restore points, the new restore point has been created. If I use the Recovery Point Browser the new restore point is shown and I can browse through it; it looks okay from that standpoint and the browser isn't doing anything to indicate a problem with the restore point.

If I look at the Backup History tab in the Tools section of the UI, the new recovery point is NOT shown, which makes sense.

Question:  Is there any way to know what portions of the aborted restore point are good?  I know there is a validation tool in the backup, and I believe I have it set to ON. Is there any other means to validate the restore points after the fact, after a reboot, and perhaps other work has been done on the client machine?

One other, perhaps fine point:  If the restore point is incomplete and invalid, why is the Recovery Point Browser so eager to show it to me without any warnings? I'm concerned that I can restore a file that is corrupted, thinking that it is actually okay.

Thanks! 

[marcogsp]

Is your USB drive NTFS or FAT32 formatted?  FAT32 destinations do not work to well with BESR imaging tasks.  See the technote below

http://seer.entsupport.symantec.com/docs/305459.htm



Here are a couple other technotes regarding causes for hanging at the juncture you describe:

http://seer.entsupport.symantec.com/docs/317111.htm

http://seer.entsupport.symantec.com/docs/307406.htm

[EHorwitz]

Thanks.

My question was really more about the validity of the (probably incomplete) restore point. I'm still looking for info on that bit.

As far as the rest, which is equally important:

The USB drive is NTFS. It is a 1TB Seagate FreeAgent drive and has plenty of room. There are 686 GB free; my original full backup was only 44GB and individual restore points since then have been 1-2GB. We have Gigabit Ethernet internally for the LAN and it's functioning nicely with no problems.

Here are my event logs:

7:00 PM:

    Info 6C8F1F65: The drive-based backup job, My Computer Backup, has been started automatically.
    Details: 
    Source: Backup Exec System Recovery

8:06 PM:

    The Volume Shadow Copy service was successfully sent a start control.

8:06 PM:

    The Volume Shadow Copy service entered the running state.

8:06 PM:

    The MS Software Shadow Copy Provider service was successfully sent a start control.

8:06 PM:

    The MS Software Shadow Copy Provider service entered the running state.

8:25 PM:

    The Volume Shadow Copy service entered the stopped state.

8:28 PM:

    The MS Software Shadow Copy Provider service entered the stopped state.

8:28 PM:

    The Symantec SymSnap VSS Provider service entered the stopped state.

(Note:  A full backup takes about 1:26 on this machine, so this is when the backup SHOULD have finished.)

11:57 PM, 7/9/2009:

    Windows cannot determine the user or computer name. (The RPC server is unavailable. ). Group Policy processing aborted.  (Note:  I have no idea what this one is about. I'm researching it now)

8:15 AM (the NEXT DAY, 13 hours AFTER backup started):

    The session setup to the Windows NT or Windows 2000 Domain Controller \\BPTSERVER for the domain BPTRESULTS is not responsive. The current RPC call from Netlogon on \\BPT02 to \\BPTSERVER has been cancelled.  (Yes, I'm looking into this one too! I believe it has to do with other events on the system that took place around the same time.)


Where else do I need to look? As my notes say, I'm not convinced either of the errors (11:57 PM and 8:15 AM) have anything at all to do with the backup issues.

I'm still concerned about the Recovery Point Browser treating the incomplete backup as if it was fine. I'd also like to know if there's any way to verify the validity of the restore point that was not finished. I did restore some files from it that I knw were changed prior to the backup and they were fine.

Thanks!

[marcogsp]

Is it safe to assume that the 1:26 time for a full backup was determined by doing a manual one-time backup?  Are manual one-time backups completing successfully?

What is the OS and service pack level on this server,?  On a Win2K domain?  Any Exchange or other mail servers running on this server?  Any database services running on this server?  If any mail or database services what type and versions?

If manual backups are completing successfully, but the scheduled jobs appear not too, then there may be a maintenance task, either disk, mail or database that is occurring  during the scheduled imaging  job.  Generally, VSS errors occur during the first 5% of the imaging job.

Wish I had a simple definitive answer regarding verifying the completeness of your images.  If this were happening in my shop, I would run the job with no compression.  Then if the image size matches the used space size closely, you could feel better about the images being complete.  Since the image is opening in the Recovery Point Browser, you could also "recompile" the image to no compression and make the same comparison. 

http://seer.entsupport.symantec.com/docs/292813.htm

The verify option within BESR only verifies that files within the image can be opened, not make a file by file comparison against the source drive. You could also mount the image as a virtual volume within the Recovery Point Browser and run any file comparison utilities against the source and virtual drive.

RPC errors? Arrrgh!  RPC is a notoriously troublesome  protocol and has been patched into oblivion.  I'm guessing that the RPC errors are not to blame for the troubles with BESR, but it is easy enough to isolate. If you skip the imaging job and stop the BESR service, does this make the RPC errors go away?

[EHorwitz]

Thanks for the help.

This is actually a desktop backup over the network, to a USB drive plugged into a Windows Server 2003 machine.  The desktop is Windows XP SP3 fully patched. The server is also fully patched. The USB drive is set up as a shared drive, but I am the only one with WRITE privileges to the drive so that nobody else can mess with it.

The 1:26 time did come from a manual backup and yes, manual backups seem to complete without error.  I modified my schedule so I could get a better look at this; I'm now backing up 2x daily, Sun -> Thursday. My last backup is also hung at 95% right now - it was supposed to complete last night.  I'm looking at logs now. That will take a bit.

I had forgot about scheduled tasks. My nightly antivirus check starts at midnite. (Symantec Antivirus Enterprise, 10.0 - not the latest engine, I think, but the virus definition files are current.)

I thought that since BESR uses the Microsoft Shadow Copy services, other processes wouldn't matter. I'm not sure I understand the relationships there. Half of the systems here are laptops and my plan was to have them start an automated backup when the users come in and plug into the network, whenever that might be. Am I likely to have these kinds of problems with the laptops, since the users will be working when the backup starts?

Somehow, I feel like I just opened a really nasty can of worms, and I'm not planning on going fishing any time soon...

Thanks again for your help!

[EHorwitz]

Here's some more info -


Right now the icon in the system tray shows "Symantec Backup Exec System Recovery Creating Recovery Point - 95%

I cannot open the BESR client. If I try, the splash screen appears, sticks around for a while, closes, and nothing else happens.  Task Manager shows a new instance of VProConsole.exe has started (in the Processes tab).  Right now, from trying several times, I have four instances running, apparently doing nothing. All of the items shown in the menu when I right-click on the system tray icon do the same thing; they spawn another VProConsole.exe process and nothing else.

The only one that does anything is the Progress and Performance menu item. When that opens, it says it is Creating Recovery Point - 2 seconds remaining - updating history.

(Is it trying to update the Archive bits on the files, close the Restore Point on the remote drive, update the internal status of BESR, or what in those two, tiny, unfinished seconds?)

I can't get into the schedule to see what was supposed to happen between Friday afternoon and Sunday night's backup. I should not have anything scheduled, and there is a restore point file (1.7Gb) on the drive for last Friday afternoon.  That should have been an automated backup.

The only cogent event log entry is this one:

Info 6C8F17E7: An automatic recovery point was not created because an earlier recovery point from the same job was still in progress. Error EBAB03F1: Operation aborted.
Details:
Source: Backup Exec System Recovery

Again, I don't know if this is a result of running trial ware, is fixed by patches since the 8.5 trialware was released, or what.

Any suggestions are MORE than welcome...

[marcogsp]

Perhaps the job and / or history files are corrupted:

http://seer.entsupport.symantec.com/docs/293307.htm

If you try the fix suggested in the above technote, document your job settings first, because this procedure wipes ou the job an history files completely

Also if you enter Info 6C8F17E7 in the search bax at:

http://www.symantec.com/business/support/overview.jsp?pid=53845

you will get several other technotes that also apply to the error you are getting.

[EHorwitz]

I followed the procedure in doc #293307 and everything was running nicey-nicey again.

Oops... he said "was..."

I was about to close this out as solved, but I just noticed that I'm once again hung at 94% complete and the BESR client won't open again.

I had changed my backup schedule to 2x daily, just to increase the number of backups and the chance of observing a failure.  Sorry, but it worked!

Backups take place at 10:00 AM and 10:00 PM. I get a popup at 10:00 AM asking if it is okay to run the backup. How nice, polite and thoughtful! (We in IT have come a long way.)When I look at what's on the backup drive, I see:

1. the full disk image from the 15th (evening) at 32GB+,
2. the incremental from the 16th (morning) at 53MB,
3. A 2K "Symantec Recovery Environment file" with the same time/date as the 16th morning's incremental backup, and
4. A 301MB backup (incremental?) from the 16th (evening) which is apparently the one that we're hung on.

Once again, nothing of interest in the event logs.

I can do the cleanup dance again and restart things, of course. The obvious question is what is causing this, and how can I keep from having to do the cleanup job every couple of days?  Will this happen on the other desktops, too?

Looking for help, hope, suggestions, etc.

Thanks!
 

[marcogsp]

I'm still thinking that there is a conflict with either VSS or another service running on the system.  You could try disabling VSS copying altogether for all the volumes you wish to image.  This would not impact BESR's ability to image the drives as it will invoke the necessary VSS components automatically.  If VSS is trying to take a historical copy of the volume while BESR is doing it's thing, it might just corrupt the BESR historical data.

As for any services causing a conflict, those can be disabled and enabled with command scripts during the Pre-imaging, Post-snap, and Post-Imaging portions of the imaging process

You can disable the tray icon as well. I don't know if it would help with the imaging process, but it might disable that okay to proceed popup.

[EHorwitz]

Okay, I'm deleting the *.pqj and *.pqh files again and I'll be rebooting shortly. I should get a full disk backup tonight and an incremental at 10:00 AM on Tuesday.

I checked, and the Volume Shadow Copy service is currently off. It is set to Manual. I've gone ahead and disabled it instead. My antivirus scan doesn't start until midnight, which gives adequate time for the backup to complete.

Any suggestions for ways to identify processes that might be conflicting with BESR?

More later...

[marcogsp]

I'm sorry, I didn't articulate what I really meant.  I was refering to disabling the making of shadow copies for each individual volume, not disabling the Volume Shadow Copy Service altogether.  Still, it might be intersting to see if disabling the service altogether has a positive effect. 

You could try scheduling Performance Monitor to begin monitoring at say 75 minutes into the backup and see what that pulls up for resource hogs.  A 1:26 duration translates to 86 minutes and 94 % of that is approx. 80 minutes.  Beyond that, it is a guess and by gory process to figure out what  service could be interfering.

[EHorwitz]

Marco:

How does one disable the making of shadow copies for each indvidiual volume?

Thanks for all your help. I cleared the files, let things run, got a full backup and one incremental, and am again hung at 94% and missed the next backup. My logs don't really show anything interesting. The logs on my server, where the USB target drive is attached, show nothing and the daily backup there (to the same USB drive, five hours later) is working fine.

I realize I don't qualify for regular support since I'm using the software on a trial basis. Would the support I'd receive as a customer include having someone connect to my machine remotely and looking around to figure this out?

I have other questions about the product and will post them in the forum. I'm about out of time for testing BESR and will be off to trying out another product that (hopefully) will perform better.

Again, thanks for your time and help.

[marcogsp]

For XP and Server 2003 -- Double-Click My Computer..........Right-Click any local volume then choose Properties............Shadow Copies tab.  From there you can enable or disable Shadow Copies fo any or all volumes.

For Vista and Server 2008  -- Right-Click Computer ..............Properties........ System Protection tab  Ability to enable and disable for volumes exists here too.

The only thing I've ever had to call Symantec on was licensing issues, so I can't speak personally abouyt the tech support.  Friends who have used tech support say that the support incident remains open until the issue is resolved.  I have one friend who refuses to purchase a support contract and pays for support on a per incident basis.  I think the charge is around $200.00 US and he says Symantec loses money every time on his incidents.  I know.....Bwaaaaah! ;-(
I don't know about the remote access capabilities, but I could ask one of my friends, or you could private message one of the forum admins and see if they know the remote access capabilities of tech support. 

Best of luck in finding a working solution, be it with BESR or something else.  Nobody I know enjoys doing DR from scratch when there are solutions to more easily backup and restore dead systems.