Thanks,
That's interesting because I just enabled device cleaning on the
libraries recently. Until then, we had been doing it manually, not
through Legato. But, we've been seeing these "device or resource busy"
and "invalid argument" problems for a long time, always sporadically. I
checked the libraries the other day, and they do not indicate that any
cleaning has taken place, so it doesn't seem as if the library auto
cleaning is the culprit in this case, and even if it was, there's no way
it would be cleaning as often as these messages have appeared --
sometimes several times per week.
There seems to be two different camps on the device cleaning issue. I
guess we decided to let the library handle it because it *SHOULD* know
when the drives REALLY need to cleaned since they communicate this info
to the library, but it was our understanding that allowing Legato to
handle this forces you to set up a cleaning interval based on elapsed
time, not what the actual drive is communicating to Legato. In other
words, Legato decides its time to clean a drive because it's been more
than x number of days since the last time it did it, and that has
nothing to do with the amount of time the drive has actually been in use
and may in fact not be necessary since they warn against excessive
cleaning. However, at least one person was of the opinion that newer
releases of NetWorker are able to query the device to determine if it
needs to be cleaned, so not sure. Our version (6.1.1) is not that smart,
and if in fact this capability does exist, I suspect it was added much
later.
Seems, though, that if the library is busy cleaning a drive that Legato
should see that the drive is in use and wait?
Just curious what low sleep values versus higher ones might mean in
terms of other problems outside the purview of drive cleaning.
George
Mark Dyer wrote:
>
> George,
>
> I think I can help answer the function of the Sleep settings. Simply put,
> it's the amount of time the Networker media management daemon (nsrmmd)
> waits after receiving a "Task complete" signal from the library before
> issuing the next command to the device.
>
> Example:
> The Eject sleep value is set to 5.
>
> One issues the command to unload a tape from a tape drive.
>
> nsrmmd first confirms the library is accepting commands, then issues the
> tape rewind command.
> Once the tape rewind command completes, nsrmmd will wait 5 seconds before
> issuing the tape eject command.
>
> I've seen problems with "read open" errors when our library was configured
> to perform automated drive cleanings. I found that the Load sleep value
> within Legato was too long and the library had time to grab a cleaning
> cartridge & slap it into an empty drive right in the middle of Legato
> loading, unloading or labeling tapes.
>
> Similarly, when using Legato to manage device cleanings, I found the
> default Cleaning delay of 60 was too short. It takes 5 minutes for my
> DLT7000 drives to perform a cleaning.
>
> --
> Note: To sign off this list, send a "signoff networker" command via email
> to listserv AT listmail.temple DOT edu or visit the list's Web site at
> http://listmail.temple.edu/archives/networker.html where you can
> also view and post messages to the list.
> =*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
--
Note: To sign off this list, send a "signoff networker" command via email
to listserv AT listmail.temple DOT edu or visit the list's Web site at
http://listmail.temple.edu/archives/networker.html where you can
also view and post messages to the list.
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
|
