SUMMARY: Backup server device problems [HP LTO]

I apologize for taking so long to send a summary to the list about
our problems with the HP LTO/SureStore hardware, the reason is simply
that we still don't know for sure that we have a working solution.
However, the situation has improved since my first email to the list,
so I will send what I have for now.


Thanks for suggestions from several people:

Stale Schwenke
Elizabeth Lee
Matt M. Morris
And a big thanks to N.N. (wants to be anonymous). He helped me with
patch diagnosis, etc. 



Our problem was:

> we have a Ultra 5 with an Antares dual lvd SCSI card and Solaris 8 set
> up as a backupserver in combination with a HP SureStore 2/20 library and
> HP Ultrium 230 tape drive. The Ultra 5 also has a fiber gigabit NIC.
> 
> The tape drive and library is hooked up to one of the SCSI channels, and
> a Seagate disk to the other.
> 
> The software we use is Legato NetWorker 6.1.
> 
> After many hours of frustration setting this up we now have it partially
> working.  That is, backing up seems to work fine for days, but when I try
> to read the tapes I will most times get SCSI and I/O errors, with some of
> the typical errors reported snipped from /var/adm/messages below:

[the rest snipped for brewity]




Solution (need more time for testing and verification before I conlude):
------------------------------------------------------------------------

* Upgrade firmware of HP SureStore 2/20 library to 1330
* Upgrade firmware of HP Ultrium 230 drive to E19V
* Use st.conf settings from HP
* If you use HP media, check with them that it's not from a bad batch


I believe that the above are the most important factors.  In addition
you should also check the basic stuff like cabling, terminators, SCSI
card drivers, OS patches and compatibility issues with the various
vendors involved.


The rest of the email is describing the process in more detail, and
is not recommended reading if you only want the pure technical stuff,
which is given above.


--


The main problem, as I have experienced it, have been to get support
and feedback from HP.  I spent months trying to find someone there who
could send me firmware upgrades and give support in general. One of
the more bizarre suggestions from someone in HP were that I would
probably be better off contacting Sun to try to get hold of the
firmware for the HP LTO drive! (The HP website only have firmware
that can be uploaded to the LTO drive from a Windows based host.
If you have no Windows server to hook up the library/drive to, you
will need to upgrade directly from tape).

Since my original firmware versions were also listed in Legato's
compatibility matrix and I was unable to get hold of firmware from
HP I spent most of my time testing, debugging and verifying all the
other possible causes in the early stages.



Having exhausted all other possible causes, while still trying to
get support from HP I finally managed to get a tape from them two
weeks ago, with what I was told was the current recommended firmware.
Version 1300 for the library, and E19V for the LTO drive.  This did
improve the situation a lot, but unfortunately the second full backup
after the upgrade failed badly, with strange and inconsistent error
messages.

I immediately reported this to HP (with the usual problems of getting
feedback) and they suggested I should upgrade the library firmware
again, as the version should really be 1330, not 1300 as they had just
sent me on tape. They also said that the last crash we had could be due
to bad tapes from HP.

[I got the information about the possibility for bad HP media several
months ago, from a non-HP source, and have been testing using mainly
IBM tapes until the latest failure.]

One or more batches of tapes made by HP is bad, and they are now replacing
them.  They have also told me that they will only give support if I use
HP media, which I find a bit strange having had the problems I've had
with both their hardware and media.


If you are in the market for LTO based drives and libraries I would
strongly recommend that you ask the vendor to supply references to
customers who have had success with the combination of hardware and
software you are intending to use.  I know there are some vendors who
will do this (unfortunately, I learned this too late).




-Sigbjørn
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers