Replacing tape drive

[michaelwild]

My Seagate DAT died. I can get exactly the same model from Seagate-Certance. Will this be a simple remove old and replace new? SCO OpenServer 5.0.4 The system still "sees" the defunct drive on bootup.

 

[edfair]

plug and go.

Ed Fair
Any advice I give is my best judgement based on my interpretation of the facts you supply. Help increase my knowledge by providing some feedback, good or bad, on any advice I have given.

 

[PHV]

Obviously the new tape drive MUST have the same SCSI Id as the dead one ...

Hope This Help
PH.

 

[jerrymei]

You better remove the old one using

mkdev tape and choosing the appropriate option and install the new one providing the correct id, bus, and host adoptor.

Or you can you use the same command and change the default tape drive with the new one. But I think the first way is better.

 

[michaelwild]

I received the new drive yesterday. Exactly the same model as the old. Instructions included for jumpers and dip switch settings for SCO OpenServer. I'll remove the old drive, set everything the same, and plug in the new one. This seems the easy, obvious thing to do. Of the course the easy and obvious doesn't always apply to UNIX.

 

[michaelwild]

Well, easy & obvious didn't work so well. I used the Informix ontape command and the system froze. Could not get alternate terminals, nothing so I had to reboot. Tried a find and cpio command and had the same result.

I will now try the mkdev tape route. This should be done in single user mode correct? I have physically removed the old drive and installed the new one. Does the old drive need to be reinstalled prior to removing it using mkdev tape?

Any further advice/caveats will be greatly appreciated.

 

[edfair]

With my 0 for 1 with you my input probably is suspect.

Should be able to leave it in. I've never had problems rebuilding with the hardware in place. and yes, single user.

Replacing the drive shouldn't cause your problems. I switch external stuff around regularly without problem.

I would be curious what /usr/adm/messages has to say about the hardware before and the hardware afterwards. And also some curiousity about what errors tar would show.

Ed Fair
Any advice I give is my best judgement based on my interpretation of the facts you supply. Help increase my knowledge by providing some feedback, good or bad, on any advice I have given.

 

[michaelwild]

I removed the old drive info using mkdev tape then installed the new drive with mkdev. Same result, system freezes when I use the Informix ontape command. I kind of shocked at this Windows like behavior from UNIX - locking up because of problems with a peripheral device.

I did search this data base and found some previous info from edfair. May 9, 2001.

mkdev tape. Remove your scsi device, regen & reboot. mkdev tape, remove the default scsi tape, regen & reboot. mkdev tape, add your tape, regen & reboot. ...

My question is what is meant by regen? relink kernel?

Any thoughts on why the system freezes when trying to access the tape drive? Could the scsi controller be bad? The controller has the tape and a CD ROM on it. I'll try to mount the CDROM and see if it's working. When UNIX boots and lists all the devices, etc. is this a confirmation that the devices are present & working properly?

Thanks again for any help.

 

[edfair]

The previous post was probably related to 3.2.4.2 that had a problem with tape support when installed from floppies.

Yes , regen is building a new kernel with new links.

Unlikely the SCSI controller went bad with a different tape. But I would probably take the tape out of the kernel and check the CD.

What does tail -20 /usr/adm/messages tell you about your tape device and SCSI controller, names and IDs?

When it is good, it is very, very good. When it is bad, it is very, very bad. Much better ratio in general.

Ed Fair
Any advice I give is my best judgement based on my interpretation of the facts you supply. Help increase my knowledge by providing some feedback, good or bad, on any advice I have given.

 

[michaelwild]

Haven't been able to get back to this - had to cover for someone else at work. My usr/admin/messages file is an executable and root doesn't have privilege to run it. Should I just change the file type or enable execute privileges?

 

[edfair]

/var/adm/messages in 5.0.5. May be the same in 5.0.4.
One of the problems with working multiple versions. And an old, decrepit mind.

Ed Fair
Any advice I give is my best judgement based on my interpretation of the facts you supply. Help increase my knowledge by providing some feedback, good or bad, on any advice I have given.

 

[michaelwild]

Problem fixed. Took the old & new drives to local UNIX shop. They figured out that two sets of jumpers needed to be jumped. That sets the unit id. Even though the drives are the same model I don't think the older drive required this.