NT server 4.0 SP6a crashing at backup time

[mucous]

Hello everyone,

I just took over a small NT 4.0(SP6a)network (approx. 16 Win98 clients) that has been experiencing troubles when running backups. I'm fairly new to the admin role and darn glad I found this forum!
I don't have all the info yet on every issue (there are many!), but the problem with the backup seems to have begun when a user from a remote office was added locally. I've been told that when he was brought onto the lan (months ago), he brought several viruses with him. The admin I'm replacing assures me that the viruses have been removed, but could not specifically tell me what the viruses were. She's not a very technical person, and actually worked with outside support and vendors to get things to the point they are today. I should say she's no longer with the company, so I'm kinda working in the dark here.
The network is functioning during office hours, but when the tape backup begins about 10:00pm, the server crashes, bringing down email (Exchange 5.5). The past admin would just come in in the morning and reboot the server so everyone could work and get their mail. There are no specific messages in the event viewer aside from a mention that the system crashed when backup began. This is a sporadic thing...doesn't happen every time backups run, more like every other or third time.
Backup is via Veritas Backup exec. Apparently the Backup exec software has been uninstalled then reinstalled in an attempt to fix the problem. The server is a Dell, and Dell support has gone through everything and hasn't come up with a solution either. I know I'm probably not providing enough info to go on, but thought maybe you may be able to offer some ideas on what to look into to begin rectifying the problem. I was thinking that I might stay after work when everyone's gone home and run a repair install of the server software to see if it might detect and repair the system files. Before the virus issue, backups went fine. After the viruses were cleaned from the system, Veritas gave the old admin a small ritual of shutting down the backup services, deleting corrupted files, then starting up the backup services again. This is the current workaround and as I said, only works sporadically. Would I be mistaken in attempting to run the NT server install to repair the O/S? I'm also thinking to check for BIOS updates.
Any advice is greatly appreciated, and I hope your holidays go well!!

I apologise for the lenghty post.

Thanks in advance!

 

[djhawthorn]

Does the backup log show anything specific?
If you run a verbose log, what is the last thing it backs up before crashing?
Does the system "hang", or does it crash going to a Blue Screen of Death?
If the latter, what is the STOP error message?
Do you have on-access virus scanning software installed?

I wouldn't recommend putting the install disk in... you're asking for trouble. If you try to install over the top of the old OS, you'll lose all your user accounts and stuff. NT repair (IMO) doesn't work either. Except in a contained (training) environment, I've never been able to get it to successfully work for me and actually fix the problems I was having. MCSE NT4/W2K

 

[mucous]

dhawthorn,

Thanks for the tips. I'll look into other options as I definitely do not want to lose any profiles. My server experience is limited so far. Reinstalling usually works for some Win98 problems I've experienced in the past. That's why I thought to do the repair on the NT server the same way. I went in this AM and the server had locked up again, I guess it isn't a crash. No BSOD. I checked the event viewer and got the following:

11/28/02 event id: 9 source: pedge
Cat: none type: error
The device \devicScsiPort1 did not respond within the timeout period

There were two of these in a row. No other errors. I wan't able to get into the Veritas software until rebooting. I guess because it couldn't find the tape drive. When I did the reboot and checked the backup record, there was no record of last night's backup at all. No errors either. I assume this is because the backup software was unable to find the tape drive.
When this happens, it locks up the Exchange server, and stops all the mail from coming in or out. Last night, I sent emails to my account and got undeliverable messages back, so I knew then that the server had locked up.
I'm pretty new as far as administering an NT server network. This is my entry into administration, from past technical support roles. I certainly appreciate any help you can offer on this problem!

 

[mucous]

I should also note that the backup performed as it should the night before.

 

[Dogers]

how odd

how about trying the following-
* leave task manager open with a high update frequency, then when it hangs, you can hopefully see a little more what was going on at the time
* get the performance monitor going and leave that open too. leave it monitoring network, hd, memory (physical and virtual) - again to see whats going on at the time when it hangs..

if it continually cant see the drive, and nothing odd is showing up in the logs, taskmanager, etc, then is there any way you can test another tape and scsi card in your server? or test your tape and scsi card in another machine?

 

[PeteH2]

I realise that this is a long shot but...

When did you last clean the heads on the tape drive ? The event viewer shows that it was expecting something back from the tape drive. If the drive could not read the tape, this could be down to:

* old tape - replace tapes every 1-3 years on rotation
* dirty heads.

If in doubt clean the heads AND buy new tapes!

Do you do a full backup every day or is it a daily incremental and a Friday full backup ?

HTH

Pete...

 

[mucous]

I'll look into those suggestions Dogers and PeteH2. Thanks for the help. I've got a few other fires to put out with this. Didn't really get an opportunity to address the backup issue, plus the last one went ok. I'm willing to bet it'll screw up tonight. Will pass on any and all error info I come across. I really appreciate your experience! Thanks again.

 

[caddnima]

For now, try to disable the backup software/hardware... Just to see if it will still hang/freeze. And if you could seperate the backup drive to another computer... like, use an old computer to make the backup so that anything happens like this, it won't affect your server...

we are a small office, and we dont use tabe backup...(not that they are not reliable but they are slow when retrieving some files...) we use hard drives since it is cheap now a days... doesnt need to be 10k rpm...

 

[PeteH2]

Caddnima, I take it you use removeable hard drives, so that you can store your back up off site, therefore overcoming the risk of fire ?

Pete...

 

[caddnima]

Yes, actually I do use the a docking bay so I can remove it once a week for offsite backup...

$20.00 for each docking bay (IDE). I have actually 3 40GB that are rotating. 1 in use, 2 offsite...

 

[PeteH2]

Good idea! and good on you for practising 'safe hex'

Pete...

 

[swissmike]

Hi everyone,

I'm having a similar problem with backup exec and an exchange server. It seems every couple of weeks I will come in in the morning and the server will be down. Looking in the event log just indicates that the previous shutdown was unexpected a few minutes after the backup should have run. No backup takes place and the server hangs. I'll be interested to see how this progresses.

 

[mucous]

For some odd reason, I haven't had any crashes this week. That's a good thing right?
I like the talk on using a removable hard drive instead of tape backup. I think I'll bring the subject up with my boss. Is it possible to run a backup from the server to a client machine? I'm logged into a win98 client as admin. Was thinking if I could run the backups to a removable HD on my win98 client, then that would just be lovely. I have a spare IDE drive big enough to backup the crashing server plus another server that's not having issues. Just need to get one of those kits to drop the drive into. As I understand it, the backups would move much more quickly also, correct?
One last question...Does Backup Exec support backing up to a remote hard drive? Is version 8.5

Thanks all. You're very helpful!

 

[caddnima]

Mviker,

It would be helpfull to put your backup to another machine. I doesn't need to be fast or hight tech as long as it is connected to the server. You can switch hard drive without affecting the server. (you have to turn off the computer to switch the hard drives from one docking bay to another)

Server is W2K. Backup is a Pentium 133 with 32MB on NT4.0. It has a fixed 40GB drive to do the weekly backups and another removable 40GB for cloning. Removable thru a docking bay - IDE.

I use xxcopy (can be downloaded for free thru

http://www.xxcopy.com).

It is a command line backup that I can put to a batch file that is being accessed thru scheduled tasks manager (built in for NT and W2K thru latest service packs).
The first time you use this it will take several hours to backup the whole hard drive then the next succeding backups will quicky.(only new files are being backup)


With regards to Win98, Im not sure if they have task manager but you can use "windows scheduler" for free at download.com

 

[bob47301]

Hi, I had a very similar experiance with a client. Had to do with the PERC controller card on the Dell server. Look to see if Veritas is writing checksums to the tape - uncheck if it is.

Good Luck,
Bob

 

[jasonboyes]

I have just a question is the exchange accessible (web client or POP3) to the web. If so check to see if someone is loggin into the the mail server about the time that the back up fails.

This: \devicScsiPort1
is that your tape drive or another scsi device on the chain?

Jason
Rich Cook -- "Programming today is a race between software engineers striving to build bigger and better idiot-proof programs, and the Universe trying to produce bigger and better idiots. So far, the Universe is winning."

 

[ntmntm]

I seem to recall Veritas having some kind of patch available in relation to backups and Exchange but I don't remember what versions it applied to. Check their support site.

 

[Gifted1]

It is backup exec causing it. You need to go their website and get the hotfix's for the version you have.