[HCoop-Discuss] [HCoop-Announce] Mire back up

[Steve Johnson]

On 2/12/08, docelic <docelic at mire.hcoop.net> wrote:
> On Tue, Feb 12, 2008 at 09:49:52AM -0800, Steve Johnson wrote:
> > I hate to be a prick about this but,  mire was down 24 hours because
> > of this?  Maybe we need to come up with a better way of dealing with
> > mire being down.
> >
> > I don't know how others feel, but having my site down for 24 hours
> > because of a config issue, is real annoying.
>
> Your site was down due to this specific issue for about 12 hours.

My mistake, it felt like 24 to me.  Maybe it came back up while I was sleeping..

> In any case, you are oversimplifying things here - it is almost always
> a matter of some kind of "config issue" here or there.

Sounded like a simple config issue to me. .But hey, what do I know,
only been a unix admin for 14 years.

> > And before you all start spatting on about how this is all volunteer,
> > please don't.. That does not matter when people are paying money every
> > month..
>
> I wouldn't bring attention to the 'volunteer' aspect, but I would bring
> attention to the non-profit aspect. While technically, you really are
> 'paying money' to HCoop, it is perfectly clear where the funds go and
> how they are used.
>
> Only with that in mind, can the sentence "when people are paying money
> every month" be given the appropriate context.

You are wrong here.  The context does no matter. If this kind of
outages continue, do you think people will continue to pay?  After all
it is a service, right?

> > Yeah, so its only $5, but to have the server down 24 hours
> > because of a simple config issue..  Not good.
>
> First, I would appreciate if our willingness to disclose any and all
> details regarding technical issues was not used back against us.

That's the nice thing about a coop, at least to the members, no cover
ups.. Being open about what happens is all part of the deal.  You
should always expect it to be used against us.

> Second, I run my company website on Mire too.

The outage did not effect you?

> Third, not long ago we posted a message on hcoop-sysadmin, asking for
> system administrator volunteers who'd like to join the team to speak
> up. No one did.

Not everyone wants to be a sysadmin.. After doing it for 10 hours a
day, I don't to come home and maintain another server farm.  I don't
even really see the point here.. Just sounds like you wanted to
complain about people not wanting to be admins.  Okay, complaint
heard.. now what?

> > How many servers do we have at hcoop?  Maybe one should be configured
> > to act as a back up web or daemon server for when mire goes down for
> > what ever reason.  That way instead of impacting the coop members
> > ability to have their pages served, it would only impact shell access.
>
> Yes, we could do something about this. Nathan Kennedy donated one
> Mire-comparable machine to HCoop, and that machine is currently sitting
> offline in our cabinet at Peer1.

I think we should move forward on getting this online.

> And we have another machine, the old Abulafia, which should be used as
> a shell server.

This is a good idea too.

> Our primary task at hand is to establish outpost.hcoop.net at the
> Cernio tech center, to use as secondary DNS and monitoring station
> on a different subnet.

Stability and redundancy should be the #1 priority.. We lived with out
the secondary DNS this long.. lets re-focus on getting a redundant web
server up instead.  This is just my opinion, I don't expect you to
agree.

> After that is taken care of, I will coordinate with Nathan to set up
> Abulafia as a shell server.

Good job!

> And after that, we might go into setting up Nathan's machine as a
> redundant server for Mire. I wouldn't use it as a "stand by" server,
> I'd put it in production by implementing simple round-robin DNS
> method or similar.

This should be a priority, not an "After that".  Web services are the
most impacted when mire goes down, and web services are probably the
most commonly used service at hcoop. (maybe mail is #1.. Web #2).. So
this really should be #1 on the list here.

-Steve

-- 
Blog - http://www.fratm.com/index.php/User:Fratm
"Knowing others is wisdom, knowing your self is Enlightenment." - Lao-Tzu