some questions

[GuidoLenk]

hi,

i have Solstice Backup 6.1 Build 186 running on an Solaris Sparc v8.

now i have a new library with LTO2 and i cannot define LTO2 just LTO. How can i tell Networker that it is an LTO2 Drive in it?

NQ: Can/SHould i update my Version?

And the urgent Problem: i testet some backup-speeds and i am quite satisfied with the speed, but now it always says device busy. i tried nsrjb -H but it doesn't do anything, or i can't see it. i event powerd off/on the library. Is there a other way to get it back ready?

Sorry for my english....

Would be nice to get some help!

Greetings
Guido

 

[GuidoLenk]

i should mention that i added the lto2 library to my already installed AIT-2 Library. i have bought new slot licences.

 

[wallace88]

Solaris 8 requires at least patch 108725-13 for LTO-2 support. Check and see if you have this patch.

Please refer to:

http://sunsolve.sun.com/pub-cgi/retrieve.pl?doc=fpatches/108725&zone_32=lto-2

According to the NetWorker V6.x compatiability guide, lto-2 is supported.

ftp://ftp.legato.com/pub/HW_Support/compatguide/NW_60/NW_60_HWCG.pdf


NetWorker 6.1.4 has just been released. You may want to upgrade to at least this level.

http://portal2.legato.com/resources/downloads/networker.cfm

 

[GuidoLenk]

hi wallace,

thx for this quick answer!

Can i update from Solstice to Original Legato without Problems?

 

[wallace88]

from an application point, you should be able to reove the Solstice version of NetWorker, and install Legato NetWorker without problems. It involves a package remove and package add. Please refer to the installation manual for complete deails:

Legato NetWorker Installation Guide, Release 6.1, Solaris Version

http://web1.legato.com/infodev/publications/NetWorker/UNIX/6.1/solig.pdf

Legato NetWorker Release Supplement, Release 6.1.4, UNIX and Linux Version

http://web1.legato.com/infodev/publications/NetWorker/UNIX/6.1.4/uxrs.pdf

First, try updating the Solaris patch and see if it resolves the issue.

Moving from Solstice to Legato NetWorker may have licensing implications.

 

[605]

Update - there is no technical problem, just a problem with
your licensing contracts.

To reset a busy device - simply kill the nsrmmd process. NW
actually monitors the device and will restart the nsrmmd process after 2 minutes (default). However, with earlier
versions, it is not easy to figure out which nsrmmd process
serves which device.

 

[GuidoLenk]

hi,

sry. i killed nsrmmd processes and they restarted.. but device is still busy

any other ideas?

 

[GuidoLenk]

btw: i checked for that LTO2 Patch, it is already installed, but in networker i only have LTO?

 

[wallace88]

Did you happen to install IBM's atape driver as well on the server?

If so, you will have to disable this. NetWorker uses the st driver.

 

[GuidoLenk]

i didn't do it by purpose, how do i find out if i did it?

what about the busy device? any idea?

 

[wallace88]

Drive shows busy could mean a lot of things...

Since you killed the nsrmmd's and problem still occuring, I suspect it's something to do below the application layer, either the device driver or at the hardware level.

Since it's been days, I'd suggest the brute force method and to reboot both the jukebox and the Sun box in the following order:

shutdown Solaris
power off jukebox
wait
power on jukebox, wait till it's booted up
power up Solaris

Did you look at /var/adm messages? Are there any messages related to your tape drives or scsi controller?

 

[GuidoLenk]

ok. i rebootet both and now its working again, but my lto2 tapes only use 20G and then they are full, even at lto1 thye should have 100GB capacity, since they are lto2 (i cannot edit the device to lto2, still no such entry) i would/should have 200GB for each tape. and my jukebox just has 24 slots of 72 occupied.

I guess i have to switch the device in networker to lto2 but i still cannot figure out how! :-(

 

[wallace88]

The fact that it only wrote 20 GB becoming full means it's either an issue with the device driver, tape drive, or bad tape.

I suspect it's a driver issue.

You still need to look at /var/adm/messages

Are you using ibm's atape driver?

If you are using the st driver, is your st.conf setup correctly? What is in that file that is NOT commented?

 

[GuidoLenk]

how do i find out if i am using ibm or st driver?

 

[GuidoLenk]

Dec 17 10:22:24 idomeneo scsi: [ID 365881 kern.info] /pci@8,700000/scsi@6,1 (glm1):
Dec 17 10:22:24 idomeneo Cmd (0x2f2d0e0) dump for Target 1 Lun 0:
Dec 17 10:22:24 idomeneo scsi: [ID 365881 kern.info] /pci@8,700000/scsi@6,1 (glm1):
Dec 17 10:22:24 idomeneo cdb=[ 0xa 0x0 0x1 0x0 0x0 0x0 ]
Dec 17 10:22:24 idomeneo scsi: [ID 365881 kern.info] /pci@8,700000/scsi@6,1 (glm1):
Dec 17 10:22:24 idomeneo pkt_flags=0x0 pkt_statistics=0x60 pkt_state=0x7
Dec 17 10:22:24 idomeneo scsi: [ID 365881 kern.info] /pci@8,700000/scsi@6,1 (glm1):
Dec 17 10:22:24 idomeneo pkt_scbp=0x0 cmd_flags=0x1860
Dec 17 10:22:24 idomeneo scsi: [ID 107833 kern.warning] WARNING: /pci@8,700000/scsi@6,1 (glm1):
Dec 17 10:22:24 idomeneo Connected command timeout for Target 1.0
Dec 17 10:22:24 idomeneo genunix: [ID 408822 kern.info] NOTICE: glm1: fault detected in device; service still available
Dec 17 10:22:24 idomeneo genunix: [ID 611667 kern.info] NOTICE: glm1: Connected command timeout for Target 1.0
Dec 17 10:22:24 idomeneo glm: [ID 280919 kern.warning] WARNING: ID[SUNWpd.glm.cmd_timeout.6017]
Dec 17 10:22:24 idomeneo scsi: [ID 107833 kern.warning] WARNING: /pci@8,700000/scsi@6,1 (glm1):
Dec 17 10:22:24 idomeneo Target 1 reducing sync. transfer rate
Dec 17 10:22:24 idomeneo glm: [ID 923092 kern.warning] WARNING: ID[SUNWpd.glm.sync_wide_backoff.6014]
Dec 17 10:22:24 idomeneo scsi: [ID 107833 kern.warning] WARNING: /pci@8,700000/scsi@6,1 (glm1):
Dec 17 10:22:24 idomeneo got SCSI bus reset
Dec 17 10:22:24 idomeneo genunix: [ID 408822 kern.info] NOTICE: glm1: fault detected in device; service still available
Dec 17 10:22:24 idomeneo genunix: [ID 611667 kern.info] NOTICE: glm1: got SCSI bus reset
Dec 17 10:22:24 idomeneo scsi: [ID 107833 kern.warning] WARNING: /pci@8,700000/scsi@6,1/st@1,0 (st8):
Dec 17 10:22:24 idomeneo SCSI transport failed: reason 'timeout': giving up
Dec 17 10:22:24 idomeneo root: [ID 702911 daemon.notice] Solstice Backup media: (notice) LTO Ultrium tape 000002L2 used 20 GB of 100 GB capacity
Dec 17 10:22:24 idomeneo root: [ID 702911 daemon.notice] Solstice Backup media: (notice) LTO Ultrium tape 000002L2 on /dev/rmt/1cbn is full
Dec 17 10:23:18 idomeneo root: [ID 702911 daemon.notice] Solstice Backup media: (notice) Volume "000002L2" on device "/dev/rmt/1cbn": Block size is 32768 bytes not 65536 bytes. Verify the device configuration. Tape positioning by record is disabled.


thats in /var/adm/messages doesn't look good, does it?

 

[wallace88]

Connected command timeout for Target 1.0

NOTICE: glm1: fault detected in device; service still available

Target 1 reducing sync. transfer rate

got SCSI bus reset

SCSI transport failed: reason 'timeout': giving up


Well.. this definitely looks bad... This indicates a problem with the controller or driver. You need to contact your hardware support or call Sun support to help you with this.

SCSI bus resets will be the death of your backups, because whel a bus reset is sent to the tape drive, it results in the tape drive rewinding itself. Unfortunately, NetWorker or any other applications does not know that this is happening and therefire data is written over. The SAN devices and controllers has to disable bus resets.

 

[LGTObug]

did you checked the cables? maybe theres an corrupt one? You could try increasing the scsi timeouts.

 

[GuidoLenk]

i checked log again. that was the only reset i have found so far. maybe it was coming from rebooting the machine / librarys?

but there was a message about blocksize... has that something to with LTO/LTO2 ??

i still cannot switch to LTO2.... i have seen st in log, is that the st-driver someone mentionend?

what else can i try to get LTO2?

Guido