Backup Exec ... There must be a better solution!!! 2

[TomRutledge]

Up until now that has always worked, but here is my current dilemma:

2 consecutive scheduled nightly backups failed. I rebooted my network (and felt the rath of all my users) strictly to cure this problem. All services loaded and it looked like things were going to be fine. I attempted to do a manual backup at that time. It has now been running (or idling) for over 24 hours and has totally locked me out af any ability to Abort; Halt; Hold; Disable; Escape or otherwise stop this job. I won't have an opportunity to reboot until about 3AM. I have tried all the ways listed to regain control of this program. All buttons, etc. are dimmed to Abort the task. I attempted to disable the tape drive in Device Manager. Nothing seems to work. Veritas doesn't seem to offer any support unless you have a Thechnical Srvice Contract. I'll bet I spend a quarter of my time babysitting failed tape backups. There must be a better way. Any thoughts?

 

[2ffat]

I've noticed this behaviour in our ver 8 also. Luckily, I have not noticed it in ver. 8.6. That doesn't mean it doesn't happen, I just haven't seen it in either of our servers running 8.6.

Time to update?????


James P. Cottingham

When a man sits with a pretty girl for an hour, it seems like a minute. But let him sit on a hot stove for a minute and it's longer than any hour. That's relativity.
[tab][tab]Albert Einstein explaining his Theory of Relativity to a group of journalists.

 

[jwlchan]

I do agree that this Veritas Backup EXec is not the best backup program in the world. I experience a lot of problem with this backup program myself.

I do not agree with the comment below at all:

When a man sits with a pretty girl for an hour, it seems like a minute. But let him sit on a hot stove for a minute and it's longer than any hour. That's relativity.

 

[moram]

Can you stop and start the BackUp Exec services?

 

[flo2980]

We also use Veritas 8.6 and when it hangs like that the only thing to do is a reboot on the server. You cannot stop the services through the task manager.

I would like to ask how big of a backup job is running? I noticed that we have one backing up about 35 gigs of data and it doesn't finish in a reasonable amount of time. It takes approximately 7 hours. Does this seem too long?

 

[TomRutledge]

I am backing up about 25 gig's. I takes about 3-1/2 hrs. Half of that is writing the data and half is verifying it. Backup Exec often fail on the verify pass. Length of time could be dictated by what kind of tape drive you have, I guess. Ours is a Sony 10000.

 

[GreenTeeth]

Hi chaps,

I have been forced into learning about Veritas, and finally got it working fairly well. I did however have your problem Tom but with Seagate back up, which is the same peice of software bought up by Veritas, so it is possible very slightly earlier (not next to it to confirm I'm afraid) However, our other server runs veritas, and bar the name at the top, I can not find a single difference otherwise.

First off,are you backing up the same stuff each day, or are you backing up say driver E on monday, F on tuesday etc.?

See which one is failing, and look to see if any log has been written at all, might give you an idea.

If you are backing up more than one drive, back up one at a time to find the one that is locking up your BE. Have a look at that drive, defrag it, scan it, is it old and full, and look at what you are trying to back up. When you find the Drive or maybe just folder or file that is causing the problem, you will have to reboot the network. I tend to arm myself with a good quantity of swearwords. Or if I am feeling particularly diplomatic, just make it known that the network will be down at a certain time, maybe lunch time.

I am still looking into this, as on our backup, it was trying to backup the system files, and open files, of which we don't have an open file system, (that is extra, it doesnot come as standard with BE) which in my mind is a waste of time, as you should have the operating system and any applications you use on CD ROM when you purchased the software, so you could of just reinstall it if the system went into melt down.

I have to admit, for us it was lucky, cos we just simply moved all data files onto a different drive.

Have a look at the veritas knowledge base, it can often point you in the right direction.

If you are using 8.0 sounds like you don't have the latest of kit and the boss won't pay for more. Don't get sucked into the idea that it should always be fixable. It should, but it only can if upgrades are made and systems looked after. I tend to take the attitude now that if it doesn't work now, no end of tampering is going to make it fail any more, so it is a good time to learn a skill, and if the boss is not concerned about his network, then that is down to him, not you.

Hope this helps some.

 

[GreenTeeth]

Have a peek at this too.

thread481-447380

 

[teee]

If you are using Windows NT/2000, you can just stop and restart the backup exec services instead of rebooting the entire server. You will find 'Services' in the control panel under administrative tools.

This is the only way I know of to abort a "stuck" job.

What version of Backup Exec are you running? What did the logs have to say about the failed jobs? or did they disappear when you rebooted?

 

[terrahawk]

flo2980, it depends on your drive. I just upgraded to an HP 160/320 SDLT drive and it backs up and verifies our 50 GB of data in about 2:20 hours (includes running the SQL agent). My older 20/40 DLT drive was taking 8 hours to back up 38 or so GB (compression really slows it down

Peter

 

[Qdog]

When you get a backup job stuck in Backup Exec, it is usually waiting for a complete command from the tape device. Absent of this command, BE will wait until it receives it. Bottomline however, you do not need to reboot your server to clear stuck/hung BE program.

1. Try stopping the BE Job Engine svc first.
2. If step 1 doesn't work, Task Manager|Processes, End Process on bengine.exe
3. If step two fails: Start|Run|drwtsn32 -p <pid#> where PID# is the PID# from step 2 for bengine.exe
4. In the extreme case where step 3 doesn't work, use the 'kill' command that comes with Win2K support tools off the Win2K Server CD. This always works. kill <pid#>
-Once you get BE Job Engine or bengine.exe stopped, then the rest of the BE svcs should stop normally.

 

[PaulC245]

We had a similar problem on our 8.6 servers. We have 40 branch offices and this was happening nightly and randomly among the 40. All have identical hardware and software configurations.

On some of our lockups even the Kill command wouldn't stop the process. The OS got locked up so bad we sometimes had to hard boot the server as it wouldn't do a proper shutdown.

We worked with Veritas tech support and one of the things they suggested was to stop the Open File Option. We did that and the problem hasn't happened in over a week. I'm still waiting for word from Veritas as to what we need to do to get OFO to work properly.

 

[TomRutledge]

Re: Stopping Services... It times out before successfully stopping any services. This was one of the first things I tried.

 

[jlta]

Tom, kill the processes. Close backup exec, stop as many services as you can and then use kill.exe to stop the processes. the backup exec processes all start with be (beengine.exe, beserver.exe, etc.) You can see them all in the task manager.

If you dont have kill.exe goto

http://www.google.com

and search for kill.exe

peace.

 

[styszka]

A better prodict would be Powerquest's V2i Protector 2.0

This product can be coupled with Backup Exec, or left on its own. It provides Virtual Volume Imaging at the sector level, is very fast, and can FULLY restore your server to last point of backup. Bare-metal Backups of your server are done in mere minutes, as well as a great image browser to restore single files, as well as directories from the image. A current setup that I monitor has 2 servers with V2i on it that image their drives to a SNAP server, and at night Backup exec copies just the images to tape so that they can be stored off-site. No open file option, exchange, or remote file agents needed. All you need is a compy of BAEXEC and V2i. These 2 programs together are cheaper than BAEXEC and the Agents. It is definatly worth a look see.
Stan
styszka@beringer.net

 

[mkenny]

We use the V2i along with backup exec 8.6. They give you a great combo. It is like ghosting raid 5. We have tested it and it works. We only use the V2i for a few critical servers taking snapshots maybe once a week. We can restore the machine and then use BE to put it back to the coreect time and date.

All in all BE has never failed me in restoring dataor a OS. The OS takes quite a few steps to get it back. But then again thats job security

 

[raquetman75]

It sounds like too good to be true I'm just kidding, but do they offer an evaluation copy? I want to test first before buying.

 

[Poco90]

I had the same problem about 6-8 months ago. Followed nearly every suggestion in the Veritas Knowledge base (and some mentioned above)and still the backup hanged. You were also unable to abort (end program, services), except for rebooting the server. Left with no other alternative, I re-installed the software and the problem disepeared?

 

[30362]

Raquetman75 - you can contact PowerQuest for an eval of the product - I tested version 1.2 and it worked quite well. You can image multiple servers to multiple network locations, or all servers to one destination. It works great with Exchange and SQL provided you stop all the services before the snap (which takes roughly 10 seconds) and then restart them.

One thing to be VERY cautious about, however, is how you do your testing. Make sure not to have any Open File Agent installed on the server before installation, and make sure that you use a test server or at least have several good backups.

I just recently tested v2.0 of V2I, and the only thing that I have been able to do with the new version of the software is to render my server (and laptop - I tried the desktop version too) totally unusable - except maybe as a paper weight. PowerQuest is aware of the issue (a bluescreen in NTOSKRNL.EXE immediately upon reboot) and are working on a fix, but so far to no avail.

1.2 works fine, and the only thing that I am told you gain from 2.0 is the ability to do incremental snapshots, which is critical for file servers, but since I am only interested in using V2i for my Exchange boxes, and will be doing a full nightly, the difference was irrelevant.

One other product to look at is Veritas Storage replicator, it acts much like V2i but seems a bit more stable and instead of creating a volume image in an alternate location that requires an application to read, it actually creates a complete replica of the source data in a second (or third, or fourth, or fifth ...) location, this was especially appealing to me since not only do I want to perform backups but I also want a &quot;hot spare&quot; of the data on a different server so that instead of making users wait in the event of an outage I can simply change the login scripts to map drive letters to different servers and users can continue to work while we rebuild hardware. In addition, it allows you to throttle the bandwidth utilization, we use this feature to create replica copies of critical systems over the wide-area for disaster recovery purposes.

Good luck.

 

[rvsharpe]

This is an update to the thread regarding v2i protector. I seem to be having problems creating image files to a network drive. The image creation completes &quot;successfully&quot;, but when I do a manual verify I receive a failure due to a &quot;corrupt or unsupported file structure&quot; message. If the file is created locally then the image verifies just fine, so make sure you verify your images after creation.

rvs