cancel
Showing results for 
Search instead for 
Did you mean: 

Frequent media errors code 84

varia_ketan
Level 2

*** Moved to new discussion from https://www-secure.symantec.com/connect/forums/veritas-netbackup-console-frequently-getting-media-wr... ****

hi, 

Just need a small clarification.

 

In the same scenario below things have been done

1. Upgraded the firmware of LTO4 tape drive to latest firmware.

2. Upgraded the HBA card of the firmware which is connected.

3. Checked media and no problem with it.

4. SAN environement is also OK.

5. Even I have provided the correct SCSI reservation for media type in netbackup.

 

But even after doing same things still we are getting frequent media errors code 84 on alternate occasion.

So my question is

1. Is there can be problem with the file level inconsistence?

2. The issue we are facing is only on AIX server, but while writting data of solaris server is not giving media write error on the tape, so what can be issue?

3. Is any further things which we can check for the resolving the issue?

1 ACCEPTED SOLUTION

Accepted Solutions

mph999
Level 6
Employee Accredited
I think you'll find the issue is something to do with the AIX server - if it's working on Solaris using the same drives. If I recall, the backup fails after a couple of hours, if the Solaris backups are always much shorter, it might be the case that it could still fail on solaris if the backup was a few hours, but, I think this is unlikely (based on the fact suspect your solaris backups run for more than a few mins). The issue as I mentioned in the other post can have multiple causes, I would imagine the most likely is the AIX tape driver, as mentioned by Jean-Pierre. As Nicolai mentioned, in this particular case it probably isn't the tape drive itself (faulty drives can cause this, but by the fact t works with Solaris ...). Probably the majority of read/ write errors, as I'm usually the first to point out are drive related, but I think you have one of the exceptions here. How the drive works is simple - NBU bptm process sends data to the buffer (memory) and from there it is set to the operating system, with a note attached saying 'please write to tape using buffer size xx'. The OS then performs the write to tape - so, as you see, NBU doesn't ever actually write anything to tape itself, which is why this probably isn't a NBU issue. I say 'probably' as there are sometimes the odd exceptions, but in this case, I have never seen NBU cause this fault.

View solution in original post

7 REPLIES 7

Nicolai
Moderator
Moderator
Partner    VIP   

2: You need to investigate the AIX errors. Basically a status 84 means "something went wrong". It does not mean the tape drive did a write error. 

Jean-Pierre_Bai
Level 4
Partner Accredited

Watch out with AIX because you also have IBM standards for tape drivers, look at which tape drivers are being used on these systems is it the IBM one or another generic one. And look at the properties. In any case if same drive is used with Solaris and AIX it means its a driver problem. Of course that also means driver properties.

varia_ketan
Level 2

Hi,

 

In setup we are using SL500 tape library which has LTO4 & LTO3 tape drive of HP vendor.

 

Is there anything we can check from HP. Is there any tool or link or docs which can proivde detail information about how the tape drive works when the backup is fired and can provide more detail in terms of troubleshooting

 

mph999
Level 6
Employee Accredited
I think you'll find the issue is something to do with the AIX server - if it's working on Solaris using the same drives. If I recall, the backup fails after a couple of hours, if the Solaris backups are always much shorter, it might be the case that it could still fail on solaris if the backup was a few hours, but, I think this is unlikely (based on the fact suspect your solaris backups run for more than a few mins). The issue as I mentioned in the other post can have multiple causes, I would imagine the most likely is the AIX tape driver, as mentioned by Jean-Pierre. As Nicolai mentioned, in this particular case it probably isn't the tape drive itself (faulty drives can cause this, but by the fact t works with Solaris ...). Probably the majority of read/ write errors, as I'm usually the first to point out are drive related, but I think you have one of the exceptions here. How the drive works is simple - NBU bptm process sends data to the buffer (memory) and from there it is set to the operating system, with a note attached saying 'please write to tape using buffer size xx'. The OS then performs the write to tape - so, as you see, NBU doesn't ever actually write anything to tape itself, which is why this probably isn't a NBU issue. I say 'probably' as there are sometimes the odd exceptions, but in this case, I have never seen NBU cause this fault.

Jean-Pierre_Bai
Level 4
Partner Accredited

Can you open a case with IBM ?

You can get information about the error from the errpt command. errpt -a

Cross check errors in errrpt command and tape errors seens in error log in /usr/openv/netbackup/db/media directory, file error.

Also look at tape drivers used by drives known to AIX. Also look at version of ATAPE drivers these are the standard IBM tape drivers.

Attached site shows AIX commands which can come in handy, if you are not familiar with them.

http://www-01.ibm.com/support/docview.wss?uid=swg21600215

 

Another thought, check you AIX server is at a software version supported by NBU.

Marianne
Level 6
Partner    VIP    Accredited Certified

For non-IBM tape drives, use AIX ost (other SCSI tape) driver.

Please have a look at NBU Device Config Guide - on AIX, variable block length must be enabled as well as extended file marks.

varia_ketan
Level 2

Thank for the update and info. I will check it out and get back to you.