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

[Nathan Kennedy]

Michael Potter wrote:
> On Tue, 2008-02-12 at 18:50 -0500, Adam Chlipala wrote:
>   
>> P.S.: I still think hiring paid staff should be pretty high on our to-do 
>> list, and this kind of incident underscores the benefits.  To me, those 
>> benefits are definitely worth an even share of a reasonable full-time 
>> sysadmin's salary.
>>     
>
> I still disagree. If we're going to spend money, let's invest in an
> external disk array, or possibly a product that will allow remote
> reboots on all the systems. 
>
> Michael Potter
For the present, I still agree with Michael.  But fundamentally I do 
acknowledge that uptime and reliability is not something that can be 
accomplished with volunteer expertise and technology alone.  Those ought 
to get us more reliability than we have now, especially as we remain 
small, and an acceptable level of reliability for most people, but at 
the end of the day there are going to be situations where to maintain 
acceptable service there has to be a human being responsible for 
responding within a short period of time.  I hope that we can do this 
with as much volunteer effort as possible, but I think at some point you 
have to compensate to provide a meaningful guarantee.  The more hardware 
you throw at the problem too, the more things there are to break and go 
wrong; even if they provide more reliability, they require more maintenance.

I have been available to go onsite most of the time since we moved to 
Peer1, but I have not always been completely reliable, and there have 
been periods of time when I have been out of town or otherwise 
unavailable.  Likewise our remote sysadmins have been very good for the 
most part about responding in a timely fashion and putting in necessary 
work, but getting complete coverage is another matter.

On another note, after I take the bar in February 2009 I cannot say 
where I will be.  There is a decent chance I will remain in NYC and be 
available on a regular basis for on-site maintenance, but there is also 
a very good chance I will be further away or in a different state--in 
which case we will either need a replacement volunteer at least as 
reliable as myself for onsite duties or to pay someone.  Of course I 
hope that by that time onsite access will be even less necessary than it 
is now, but it is never unnecessary.  Especially if we have more 
hardware and services provided.

-ntk