Tape Engine Problems

[ankle]

I am running Build 3200 which as far as I can tell is 11.1, SP2. I am encountering Tape engine errors. The backup will run OK once and then The next time it runs I get SCSI errors and Event log errors such as this:

Faulting application TAPEENG.EXE, version 11.1.3200.0, faulting module TAPEENG.DLL, version 11.1.3200.0, fault address 0x0007cb68.

Also, if the software will sometimes see the tape as offline after the errors start. If I reconfigure the tape device, then I will usually get one good backup, but then it will fail when it starts the next one.

Let me state that I am using a Dell PV120T autoloader. It was a recent replacement for another PV120T that was working OK until it mechanically failed. It would not load tapes, so Dell sent me a replacement.

This replacement has a newer firmware than the one that is listed on CA's website. The previous device had the exact firmware that is listed as compatible.

Could that be the problem?

Also, I read in another post to disable the tape drive in the OS Device Manager. I have not done that. Does this need to be done. I am running Windows 2003 Server.

Thanks,

 

[davidmichel]

Yes the tape drives and library should be disabled from within the OS Device Manager.

 

[ankle]

davidmichel:

Well, since I posted this, I disabled the tape drive and rebooted. I then read about disabling the library. When I did that and tried to reboot, the Win 2003 server would not boot. I shut down and rebooted at least 3 times. I did not get a bluescreen, it would just never complete the boot. I shut it down and disconnected the tape device and rebooted OK. I then shut it down again and reconnected the tape device and now the OS sees the tape device and library as enabled.

I read in

http://supportconnect.ca.com/sc/kb/...&docid=342813&bypass=yes&fromscreen=kbresults

about possible bluescreen if the library is disabled, so I assume my boot problem resulted from that.

So do you think I can just disable the tape device, but not the library?

After my Wednesday night backup job failed with the SCSI errors and locking up the drive, I stopped everything and reconfigured the tape device and then ran a test job yesterday that was 28 GB and it ran fine.

However, when the nightly backup tried to run last night, it failed again with the same errors. It is just strange that it will work sometimes but not all.

 

[davidmichel]

Read through that doc, guess that could be but I have not come across a case were disabling the library caused the system to not load but I don't have a problem with leaving it enabled for now.

In general newer tape drive firmware is fine.

What SCSI controller and tape drives are being used?

 

[ankle]

The SCSI controller is an Adaptec 39160 and the tape drive is a Dell PV120T which is actually a Sony device.

I changed SCSI cables Monday. The switched cable is shorter. I hoped that would fix it, but it did not. Backup failed again.

I opened the server and reseated the SCSI card and will try again, but I am getting frustrated.

Can a job possibly get corrupted and give you SCSI write errors? When I created a one-time backup job it ran OK and backed up 28 GB with no errors. When I run the GFS job that was created when I had the previous device PV120T that failed (would not load tapes), I get the errors.

Do I need to create a new GFS job??

Thanks

 

[ankle]

It is now Friday. After reseating the SCSI card on Wednesday, the job ran OK that night. However,the Thursday evening job gave the same SCSI errors.

Is this problem V11.1 specific? I did not have these problems until I went from 11.0 to 11.1 SP2 and after the tape device was replaced with an identical one with newer firmware.

Will an upgrade to 11.5 fix the problem?

Should I remove 11.1 and go back to 11.0?

Again, it will run one job OK and then the next will not run until I shut everything down, reconfigure the tape device, etc, etc.

Please give me some more suggestions. I have to get this going soon. I am getting behind with normal work.

 

[ankle]

Although my weekend backup backed up 85% of the files, it still failed and locked up the tape device. I have to shutdown the server to stop the job, etc.

However, I think I finally have a solution to the problem. Tonight will hopefully be the test. In another forum, I found someone stating that if the 39160 is using the 39160/39160D driver, then you will have sleepless nights and trouble, which I have indeed been experiencing. The person stated to use the 39160 Ultra 160 driver. He also stated that the 39160/39160D was a Microsoft default driver.

I checked my selection, and I was using the 39160/39160D which is the default selection.

I changed it to the 39160 Ultra 160 driver and the backup ran OK last night with no errors. If it runs tonight, it will be the first time that it has run two nights in a row without locking up.

Hopefully all is fixed. I will post results.

 

[ankle]

Well, it is now 3 days later, with no better solutions. I have even changed SCSI cards to a 29160 hoping that that might be the solution. It has not helped. I am still getting errors after a job runs OK the first time. The next time it fails with SCSI errors, unless I shut down the server completely and reconfigure the tape device. If you do that, then then next job runs OK.

I think I may just uninstall and reinstall ARCserve. I may even upgrade to 11.5, but I don't know if that will fix it or not.

This is very frustrating!!!!!! Can someone help?

 

[Kracker1]

If you have the opportunity I would call CA and request the 11.5 upgrade. It should be cvered in your License agrrement.
There are also some problems with the 2003 OS. I have run into problems on 2 occassions with the Tape drive not playing nice. I used a Registry fix from Microsofts web site. See below.

http://support.microsoft.com/default.aspx?scid=kb;en-us;842411

 

[ankle]

Thanks Kracker1.

Sorry to say, I let my support contract expire. I may still buy an upgrade, but until then I need to be backing up. I think I have already looked at the registry fix, but I will recheck it.

 

[davidmichel]

usually newer firmware is not a problem.

The error you listed is an indication that the Tape Engine generated an exception &/or crashed.

I would not recommend going back to 11.0, but the reinstallation is worth a try.

 

[ankle]

I may have finally resolved the problem. I found an older post where a similar problem was resolved by changing the speed of the SCSI bus to the maximum speed of the tape device. Although during boot, the Adaptec 29160 card sees the tape device as speed of 40 MB/s and displays onscreen as such, I decided I would try this. I went into SCSI select for the Adaptec 29160 card and changed the max speed from 160 to 40. So far I have been able to run 3 jobs without having to shutdown the server, etc.

I will let you know if jobs continue to run.

 

[ankle]

davidmichel:

OK, I spent Saturday morning, reinstalling everything. I unistalled and then installed v11.0, upgraded 11.0 to 11.1 and then patched it to 11.1 SP2. I disabled the OS drivers for the tape device and the library controller as the device is an autoloader. Previously when I disabled the library contoller the system would not boot. It booted OK this time. I also created new GFS scheduled jobs after I reinstalled. I ran backups OK, with no SCSI errors during the job.

However, after a job has run, I cannot dismount the tapes from the auto loader without rebooting the server. Arcserve display will show the tapes spinning which usually shows that the device is starting a dismount. However, the tape device will not do anything and the tapes just continuously spin on the display. I can stop the tape manager, but if I try to start it back up I get errors. The only way to get out of it is to reboot. After that, I can disomount, change tapes, and remount with no problems. This has happened at least three times.

It is as if there is something not closing out after a backup job completes that keeps the tape manager from working properly.

One question that I have is that when I upgrade from 11.1 to 11.1 SP2, do I need to install a device upgrade or is the latest device upgrade included in SP2. That was not clear to me on the CA website.

As this is getting long, let me know if I need to start a new post.

Thanks,

 

[ankle]

At this time, I cannot risk doing a scheduled backup unless I shutdown and restart the server after a backup completes. It appears that this is the only way to sucessfully have it run.

I plan to upgrade to 11.5 and hopefully this will fix this problem. If it does not, then I will probably purchase different software and possibly hardware. I have wasted too much time trying to get it working properly.

Thanks for suggestions

 

[ankle]

I have not been back in awhile as I have been trying to resolve the problem, but finally decided to get the 11.5 update. I received it and I plan to update tomorrow. I would like to know if an 11.1 GFS job needs to be recreated or should it run OK. I do hope this SCSI problem goes away for good.

 

[davidmichel]

the old jobs should be fine.

 

[ankle]

I have not been back in awhile as I have still been trying to figure this thing out and make sure that I have not left something out. I have updated to 11.5 and applied latest SP and patches.

All I can say is that my backups will work fine as long as I shutdown, not just reboot, my server and restart between backup jobs.

I plan to contact both CA and Dell about this this week, but I don't expect much help.

 

[ankle]

I am in the process of getting help with CA. Was told to do a repair install on 11.5. Still havng problems unless server is shutdown after each job. If problem gets resolved, will let you know.

 

[vschumpy]

To me this looks like some sort of hardware or connection issue, especially as it needs a hard reset to clear it.

 

[ankle]

I have not been back lately as I have been talking with both Dell and CA. CA have more or less decided that this is a hardware problem and closed the issue. So, I contacted Dell and they indeed sent me a replacement tape device, magazine, terminator and SCSI cable.

Guess what ?? I still have the same problem. Everything will work fine as long as I shutdown the server between backup jobs. I have found out that after a backup job completes, sometimes I can successfully dismount the magazine by turning the device on and off. However, backups will not complete without SCSI errors unless server is shutdown first. Since ARCserve controls the device, it seems like something is not resetting properly until the shutdown.

I guess I will have to get back with CA about this... Still hoping someone will resolve this for me.