We have been using CDI to gather TapeAlert flags from drives, and in
particular the "Clean now" flag, in order to enable on-demand cleaning of
drives, instead of traditional periodic cleaning.
As frequent readers of this list already know, we have recently added a
new storage node, that shares the same jukebox as our main networker
server. The main networker server is running solaris, and our storage node
is running linux.
The problem is that CDI is working just great for the devices that are
connected to the solaris host (since networker 7.2.1 that is, it worked
intermittently in previous releases), but the drives attached to the linux
host never see TapeAlert flags, and thus networker never cleans them until
the cleaning period has expired.
When investigating this, I found that the Administrators guide briefly
mentioned CDI and what it was for, but all in all, the information
provided wasn't as specific and detailed as I had hoped for. Later on, I
found a HP presentation about networker 7, that mentioned CDI in a few
slides. From there I learned that there are several CDI-diagnostic
utilities shipped with the networker packages, along with the CDI library,
libcdi.
Bu using one of the utilities, cdi_ta, which is used to display the
available flags and their status, I was able to see that CDI is indeed
working OS-wise, and that the drive in question was indeed flagged for
cleaning - see below:
sn01:~ # cdi_ta -f /dev/nst1 -v
CDI_GET_VERSION returns 2
CDI Get_TapeAlert returns:
Tape Critical flags:
Media is clear
Read failure is clear
Write failure is clear
Write protect is clear
Recoverable snapped tape is clear
Unrecoverable snapped tape is clear
Forced eject is clear
Clean now is SET
Expired cleaning media is clear
Invalid cleaning tape is clear
Hardware a is clear
Hardware b is clear
Eject media is clear
Predictive failure is clear
Tape system area write failure is clear
Tape system area read failure is clear
No start of data is clear
When looking in the networker administrator GUI, I verified that CDI was
set to SCSI commands. Clean now was also set to No, and the TapeAlert
information windows in the tape device properties were all blank.
This leads me to believe that nsrmmd has a bug which prevents it to read
the CDI status from the drives, rather than that the problem was related
to the OS.
So, as usual, has anyone else seen this? Can it be fixed, or is there a
fix already?
If it matters, we're running SuSE EL 9 SP 2, and the following kernel:
sn01:/nsr/logs # uname -a
Linux sn01 2.6.5-7.201-smp #1 SMP Thu Aug 25 06:20:45 UTC 2005 x86_64
x86_64 x86_64 GNU/Linux
And yes, I already know that networker storage nodes aren't supported on
x64 linux platforms, but I'm ignorant enough not to care, since it should
be. :P I mean, if they bothered to support and develop for the IA64
platform, well, go figure. But that's another discussion.
//Oscar
To sign off this list, send email to listserv AT listserv.temple DOT edu and type
"signoff networker" in the
body of the email. Please write to networker-request AT listserv.temple DOT edu
if you have any problems
wit this list. You can access the archives at
http://listserv.temple.edu/archives/networker.html or
via RSS at http://listserv.temple.edu/cgi-bin/wa?RSS&L=NETWORKER
|