media errors

[Martin999]

Hi guys,

Can I have your advice :

Using Networker 7.1.3 on HPUX / IBM silos / LTO drives.

We back up a lot of data daily (about 50TB) and therefore see a number of media errors :

eg.

07/16/05 22:58:06 nsrd: media warning: verification of volume "BB0210L1", volid 3648939265 failed, can not read record 4522 of file 2 on LTO Ultrium tape BB0210L

Now the message above is fairly clear, to many of those within a given time peroiod and the tape will be removed from further use.

However, with the following: (have removed serer name and just put ... , also, all occured, in this case on same tape drive)

BB0210L1 10/08/05 20:56:55 nsrd: media warning: ... writing: I/O error, at file 4 record 13528
BB0210L1 10/08/05 20:58:48 nsrd: media warning: rd=smithers-vcs.vfl.vodafone:/dev/vf/CCPTAPE5 reading: fsr 13522 read: I/O error
BB0210L1 10/08/05 20:58:49 nsrd: media warning: ... reading: I/O error
BB0210L1 10/08/05 20:58:49 nsrd: media warning: ... reading: I/O error
BB0210L1 10/08/05 20:58:49 nsrd: media warning: ... reading: I/O error
BB0210L1 10/08/05 20:58:49 nsrd: media warning: ... reading: I/O error
BB0210L1 10/08/05 20:58:49 nsrd: media warning: ... reading: I/O error
BB0210L1 10/08/05 20:58:50 nsrd: media warning: ... reading: I/O error

Now the Legato message "reading I/O error" is not that helpful, but how many of these messages would you guys say is reasonable ???

My understanding is that to write to a tape, a high signal is required, whereas it is easier for a drive to read from a tape, and only perhaps as low as 30% of maximum signal is required, therefore, apparently (accordinf to a technical expert from a well known tape manufacturer) read errors are usually down to a tape as opposed to drive issue.

Thanks for your time,

Martin

 

[Martin999]

... one thing guys, I've just niticed the tape relabelled itself, that would account for some of the read errors ...

but any thoughts would still be welcome.

Thanks,

Martin

 

[hoinz]

Hi Martin,

I can't believe that read errors are usually down to a tape as opposed to drive issue.
During the last few month we often had the problem that there were no errors during a backup, but at the end we got a 'Media verification failed', because the drive was not able to read the tape label once again. The label could be read by other drives though.
And we also saw that tapes were relabelled, obviously for this reason.
The percentage of unsuccessful media verifications and read errors increased continuously, cleaning tapes didn't help much, and in the end we replaced a few drives in the libraries, and the problems were gone.

Unfortunately I haven't got a number, how many error messages are reasonable.
I think it depends; if you have got a library with a lot of drives, you may afford a lot of errors from one drive for some time. If you have just one drive, these errors will soon make most of your backups unusable, or good backups will take a long time, and you should replace the drive asap.

regards

 

[Martin999]

indeed, food for thought - re. read errors, I did say usually ...

I will try and get more info, it may be that they were refrring to erors at a 'bit' level, as opposed to what Legato reports as a read error.

Martin

 

[1018DADA]

In case of LTO, the chip records number of times its been used, also some networker utilities such as tapeexcer can be used to diagnose a problem along with drive manufacturers own utilities.
Also inproperly configured devices might trough read errors or could be lose or bad cables cards etc .