[9034] in Athena Bugs
Re: Incorrect reporting
daemon@ATHENA.MIT.EDU (Chris VanHaren)
Mon Mar 2 14:15:42 1992
To: paladin@Athena.MIT.EDU
Cc: bugs@MIT.EDU, bug-cview@MIT.EDU
In-Reply-To: [9031] in bugs
From: Chris VanHaren <vanharen@MIT.EDU>
Date: Mon, 02 Mar 92 14:15:25 GMT
Thanks for the note.
I am aware that the RS/6000 reporting is inaccurate, however, until the
next release comes out, this situtation is not likely to improve.
Cview relies on SNMP (Simple Network Management Protocol) to collect its
data. Unfortunately, the RS/6000 implementation of SNMP is not the same
as the rest of the Athena platforms -- on the VAX, DECstation, and RT's,
we are using our own implementation, whereas we are relying on the
vendor (IBM) version on the RS/6000. The upshot of this is that there
isn't quite all the information there that we want, and until the next
release, when we can get our implementation installed, the information
they report will be different.
Because of this, I had to hack up a work-around for cview. It now goes
around and actually fingers each of the RS/6000's in turn, and dumps the
information to a file. Cview then uses this information to report on
the number of free and in-use RS/6000's. However, since finger is slow,
and is prone to hanging if a machine is down, the RS/6000 information is
often out-of-date.
At best, any of the information that you see in cview is a close
approximation of the real world, but in the case of the RS/6000's, it is
often not nearly as good as the information on the other machine types.
I hope this helps explain the problem, and I hope that the
misinformation hasn't caused any inconvenience. Feel free to let me
know when you think the information in cview is incorrect. The cview
software certainly isn't bug-free and sometimes needs to be restarted by
hand, but I try to keep and eye on it when possible.
At least it's good to know that someone is even looking at the data from
cview.
-Chris.
