[NTLUG:Discuss] why does the mail server stop working over and over again

[Rick Cook]

On Tuesday 03 May 2005 11:17, MadHat wrote:
> On May 3, 2005, at 9:48 AM, Chris Cox wrote:
> > Jack Snodgrass wrote:
> >> They mail server seems to stop working for multiple days once or
> >> more pre month. What's worng with it? Is there one person that is
> >> too busy to notice that it's broke ( again )
> >> Is it a hardware issue
> >> Is it a software issue There is a bottleneck here someplace.
> >
> > Not sure.  I guess what we need is somebody with some time
> > to look into this.  Send me an email at ccox at ntlug.org if
> > you believe you are the person to admin the server.
>
> Spam and viruses are a big problem.  The mail server gets hammered all
> the time and gets backed up trying to deal with it all.

We also have a large number of undeliverable addresses still on the 
mailing list that are not "auto-delisted" very well. Every so often, the 
smtp-server attempts to deliver some portion of that list of "frozen" 
messages. Some of those email addresses are for currently nonexistent 
domains add still more delay cycles in the process.

>
> >> This makes linux look bad when the NTLUG doesn't even have a decent
> >> mail server. I won't even mention the inability to provide coverage
> >> via a news group. The *News Groups (Currently broken on our server)
> >> *message has been on the ntlug home page for weeks.
> >
> > We do have moderation on our lists and sometimes it means
> > that they aren't monitored and things will queue up a bit.
> >
> > I personally would like to see the newgroups return.  It's
> > probably a simple matter to make them work... while many
> > know how to put up a newserver, not too many appear to
> > know how to establish the linkage between news and mail
> > (post to news -> list, post to list -> news).  So if this
> > is your forte... again, send me an email at ccox at ntlug.org.

We have a "replacement server" sitting at my house waiting to be made 
available. The replacement process started when SuSE 9.1 was fairly new. 
The intent is to upgrade it to SuSE 9.3 and get the services working on 
it, mirror the appropriate data, and then negotiate a swap process with 
thePlanet. If they will give us two network connections at their site for 
a couple of hours, I'll cart the replacement server to thePlanet and do a 
new data mirror and do a relatively hop swap out. Otherwise, we'll have 
an hour or so of downtime some weekend.


>
> The problem with the news groups where weird and would need a fresh
> start.  Last time looked it worked for a week then stopped.  Nothing
> made sense, but this was on a older version of mailman.

As a "stop-gap" waiting for the replacement server, we could do a 
relatively high probability of success "dist-upgrade" on the existing 
server (it's running Debian woody) - that's how it got to woody.

>
> But I am glad to know that NTLUG is the flagship example of Linux.
> heheh...
>
> >> We can fix this if the powers that be will let us.
> >
> > Uh... done.. I think... if you're saying you want to
> > be our full time admin?
>
> I can't seem to get into the server after the last crash we had, or I
> could take a look again.
>
> I have found that most people in Technology have been extremely swamped
> as of this year, more so than I remember before.  Not sure why, other
> than companies are taking on more work loads, but not willing to hire
> more staff causing the already over worked IT and security staff to be
> more stressed and over worked.
>

A my System Engineering/Program Management workload is why the "some 
weekend" swap of the replacement server has not taken place.


Rick