lil - Internet Troubleshooting

[sean dreilinger]

Marc Davis wrote:
> This isn't Linux related, but I was hoping some of you might have

i'll relate it :-) and x-post to web4lib (which you already know is a
great non-platform-specific list for web-related discussion of
librarians & info pros.)

> "When you can't get access to online resources, is the problem your
> computer, your connection, the Internet, or your online vendor?  And what
> do those error messages you see on the screen mean anyway?"

web or network administrators should not learn of failed network
services (the connection/internet part above) from end users. if the
library IT staff is monitoring the network well, the staff can make
confident judgements about end-user calls/questions -- knowing the
network is performing, staff can focus on the user's question without
that little alarm going off inside (what if the real problem is that the
dns/http/smtp/route-to-isp/etc. server just died!?!).

> (a) as system administrators or technical persons, what information would
> you like to get from your on-floor staff concerning database connection
> problems or other internet related problems?  I'm sure we've all had the

if a service goes down, a staff person on-call should receive an alert
by pager if possible, and all the relevant staff should get an email
heads-up from the monitoring service.

> "ABI-Inform doesn't work!" call that tells us nothing.  What basic

if you are confident that all the network services necessary to support
abi/inform ARE working, then "ABI-Inform doesn't work!" is informative,
and you can start helping the user immediately. 

> information would you like to get?  How much troubleshooting/diagnosis do
> you think staff should/could/would be capable of doing?

reading a web-based report of network service status -- if your staff
knows an important network service is down before users begin calling in
with "ABI-Inform doesn't work!", that's great -- whoever takes the
call/question can respond intelligently (``abi/inform is fine, we're
fixing a glitch with our XYZ and will contact you when the network comes
back, ETA 2.003 minutes'') -- or the users in a private company/net can
pull up the self-service network monitor for their own
enlightenment/amusement and reduce the help desk calls.
 
> (b) what are the most confusing/troublesome error conditions that your
> staff encounter?  Here, it would have to be occasional corruption of the
> local DNS server . . . producing all sorts of fun.

cable internet outages, corruption of the cable internet provider's DHCP
database, internet weather.
 
> (c) if you were assigned the above topic, either as presenter or auditor,
> what would be key items that you would want to cover/learn about?

would turn it into an opportunity to propose network monitoring and
web-based network status reporting :-)

these tools are worth checking out if you have ``mission critical''
network services that require human attention should they ever cease to
work:

SNMP (simple network management protocol): 
     http://ucd-snmp.ucdavis.edu/

MON  (Service Monitoring Daemon)
     http://www.kernel.org/software/mon/
     other network monitoring applications:     
     http://freshmeat.net/appindex/console/monitoring.html

MRTG (multi router traffic grapher) - observe service trends.
     http://www.mrtg.org/

hope you will share your presentation with the lists!

--
mailto:sean at savvysearch.com                sean dreilinger, mlis
 http://www.savvysearch.com                http://durak.org/sean