Help with read open error, I/O error from Legato 6

[jaSun]

Help!
I am rather new to Legato and am having "issues". Please forgive my lack of knowlege as I try to explain our problem! We are running Networker 6.0 on Solaris 2.6 (on a Ultra-2) directly connected (diff-SCSI) to a Sun L1800/4 tape library.

When we first installed the juke, we were able to correctly load about 28 tapes and Legato successfully read the labels. On tape #29 (and every tape since then....we have tried MANY) we got the following error:

Apr 24 15:18:00 <server name> root: NetWorker media: (warning) /dev/rmt/1cbn opening: I/O error

From that point on, every time we try to do anything, we pretty much get the same error:

# nsrjb -H
# nsrjb -EI
# nsrjb -EI -S31
Apr 24 15:18:00 <servername> root: NetWorker media: (warning) /dev/rmt/1cbn openining: I/O error
Apr 24 15:20:01 <servername> root: NetWorker media: (warning) /dev/rmt/1cbn openining: I/O error
failed unloading drive `/dev/rmt/1cbn' to slot '31', error '69'

When the nsrjb -EI -S31 command above was run, the jukebox grabbed the tape out of slot 31, loaded it into the top (of four) drive, the drive lights come on like it is trying to do a read, but then tries to actually mount the tape. After a couple of minutes, the error is written to /var/adm/messages and to the console.

This is what always shows up in the nwadmin Messages screen:
Thu 19:11:55 media warning: /dev/rmt/1cbn opening: I/O error
Thu 19:11:55 /dev/rmt/1cbn read open error, I/O error
Thu 19:12:18 /dev/rmt/1cbn Eject operation in progress
Thu 19:14:18 media warning: /dev/rmt/1cbn opening: I/O error
Thu 19:14:19 /dev/rmt/1cbn read open error, I/O error


So far, the majority of our trouble shooting has focused on hardware. We have replaced the diff-SCSI cards in the server, replaced the SCSI cables, replaced the system and controller boards in the jukebox and uprgraded the juke's firmware to the latest and greatest level. Also, this error is thrown out no matter which drive of the four (Quantum DLT700) are chosen by Legato. Last night my compadre reloaded Legato 6.0 server and rebuilt the nsrjb.res file.

This is beginning to trouble us because the only errors that we see are the ones listed above. There are no SCSI or server errors being logged other than those produced by Legato. Has anyone else had similiar issues or possibly have any insight? Any help is greatly appreciated!!

Thanks,
Jason X-)

 

[Voyager1]

I am unsure of the Legato interface on Sun (We operate on NT), but I get the same problems from time to time. I asked a StorageTek engineer and he pointed to the Quantum DLT7000 drives (which is why I'm responding).

For each of the drives, under &quot;Configuration&quot; (you have to turn on hidden attributes in NT) there is a setting for &quot;Consecutive Errors&quot; which by default is set to 20. According to this engineer, once that threshold is passed, you will get I/O errors until it is reset. The only way to reset it is to reboot the server.

We set this &quot;consecutive errors&quot; threshold up to 200 and the problem has about 95% gone away.

BTW - his &quot;fix&quot; for the problem is to upgrade to DLT8000 Drives ($$$$) - go figure.

Hope this helps... - Bill

&quot;You can get anything you want out of life, if you'll just help enough other people get what they want&quot; - Zig Ziglar

 

[x2shy]

Have you had a look at the drive? Sometimes the leader head (I think that is what it called.) gets messed up and so it can't mount or dismount any tapes. Get your hardware vendor to have a look at the drive. I/O errors like this seem to refer to a hardware error.

 

[Voyager1]

That was my first thought as well, however we have had StorageTek come out several times and there is nothing wrong with the tapes or the drives. I have seen the leader error you point to before, but that is relatively easy to check, and tends to be very isolated. - Bill

&quot;You can get anything you want out of life, if you'll just help enough other people get what they want&quot; - Zig Ziglar

 

[jaSun]

x2shy,
Unfortunately, like Voyager1, we have had the drives checked and they appear to be good. We had suspected that (very remote chance) a bad tape had been loaded into all of the 4 drives and hosed them up. As a matter of fact, Sun came out and replaced 1/2 of the darn jukebox!

We are going to open a ticket w/ Legato today and see what they have to say because we are out of ideas. I will keep you guys posted when/if we make some progress. THANKS for your help!

- jason

 

[x2shy]

gudluk

 

[jaSun]

OK, it seems that we have earned the coveted &quot;DOH!!&quot; award. Our problem became clear yesterday morning as we were troubleshooting the SCSI bus:
As I stated above, there are four drives in our jukebox. We noticed that when the jukebox loaded a tape into drive 1, Legato started reporting errors on drive 4 <smack myself on the forehead while saying &quot;duh!&quot;> The reason we were receiving Read errors and I/O errors is because Legato was trying to mount/unmount/eject tapes from drive 0, which the system saw as drive 2, and so forth.

Here is how we fixed it:
1.) switched the controller cables (2 of them) around on the back of the jukebox
2.) did a boot -r to reconfigure the /dev and /devices directories (NT folks: This is basically how you make a Solaris box recognize new devices)
3.) re-jbconfiged the juke.

Thanks for the replies. We will be displaying our Homer &quot;DOH!&quot; award shortly. s-)

 

[collective7]

OK, but what about if you cannot reboot the box?


Any ideas?

 

[cyberted]

Hi JaSun,

you need to correct your timer for the autochanger in Legato Networker for mount/unmount!

the load / unload sleep ,must be around 40 sec. (for LTO).

Ask your Support or library vendor about this timers.

bye

cyberted