Re: Release 1.0 trouble

John P. Rouillard (rouilj@cs.umb.edu)
Sat, 27 Aug 1994 14:15:27 -0400

In message <9408271451.AA22656@data>,
Juergen Schoenwaelder writes:

>John> Printing (actually postscript generation) is broken for me.
>John> When I pull up print or postscript, I get the error:
>
>John> Failed to write <filename>: unknown option "-outline"
>
>John> I have no clue which outline is causing the problem, any
>John> ideas?
>
>The diff below will fix this one.

And it does. I can once again print my netowrk maps for my monday
meeting 8-). Thanks.

>John> Also, how do I set the flash color and timing threshold on the
>John> ip-monitor/check rechability. Right now it is flashing white and
>John> black, about 10 seconds after the inital unreachable message comes
>John> up. Is there some easy way to set the color back to red, and have it
>John> flash as soon as the unreachable message comes up?
>
>I thought the initial delay would be a feature (we have lots of
>temporary unreachables here), but you are right that this could be

Yeah, it is a nice feature, I like it for some of my remote monitoring
where the node goes unreachable due to line congestion, but for local
hosts, I would like to turn off the delay.

>handled by setting the ICMP parameters accordingly. I will make a more
>complete patch next week that will fix this and some other small
>problems.

Thanks.

I figured out the cause of the event monitor failure in ev_filter
the following commands are done:

set ip [lindex [dns a $host] 0]
set date [getdate $clock]

In my case, the dns is failing with an "Error : domainname error", and
the script is never executing the set date or anything else. This is
reasonable since box7.iddis.com isn't yet in the dns, I do host file
lookups first followed by dns, so as long as it is in the host table
everything else works.

set ip [lindex [catch {dns a $host}] 0]

works to supress the error, but it makes for an ugly response.
Something like unknown should probably be assigned to the
ip address.

Also later in the event code, when reporting local events, you have:

ined append $filter($id,window) \
"$host $facility.$level $process\[$pid\] ($date)\n"

but for a file facility and level aren't assigned (and pid may not be
assigned either) and they make a wierd looking output:

box7.iddis.com . sudo[] (Sat Aug 27 14:05:13 1994)
rouilj : PWD=/export/spool/build2/tkined/scotty_tkined/scotty/tkined
; COMMAND=/usr/bin/ls

with just that . sitting there.

Also I have another problem with the event monitor, when you try to
disconnect the monitor from a file, it hangs.

In the "modify monitor jobs" for snmp/ip monitor is there some
way to get the following:

1) the list boxes that come up have the node numbers rather than
names of the machnes, or machine variable names. This
makes it difficult to select the right entry when a number
of monitors are in process.

2) I often select 20 or so hosts to check reachability on, and
I have no way of reveiwing the names (or node numbers)
of all of the hosts. I get the first 6 or so followed
by ... and that's it. I really need some way to
read the entire list of monitored hosts.

Also, after loosing reachability, the reachability monitor leaves the
label on the icon reporting the rtt time rather than the name of the
box as it was before the reachability was lost. Also, I miss having
the name of the box in the unreachable message, since I have to guess
which box it is that is being reported as unreachable.

-- John
John Rouillard

Senior Systems Administrator IDD Information Services
rouilj@dstar.iddis.com Waltham, MA (617) 890-1576 x225

Senior Systems Consultant (SERL Project) University of Massachusetts at Boston
rouilj@cs.umb.edu (preferred) Boston, MA, (617) 287-6480
===============================================================================
My employers don't acknowledge my existence much less my opinions.