[SGVLUG] Website / Mailing List Updates, Ideas

[James McDuffie]

Thanks for the offer, I will keep it in mind as I try to come up with 
something orderly out of the functioning anarchy that binds us to this 
user group.

On 04/18/2012 08:34 PM, Dan Buthusiem wrote:
> Ah. Yes, that's what I had in mind initially, but I also have a web
> host I'm not really using much. I have the "1and1 Unlimited" plan at
> 1and1.com. If it looks like it might be useful at all, I'd be willing
> to host the site or whatever part on it. I've had my account there for
> a few years now, and I'm not looking to go anywhere. Also, it's hosted
> with Linux. ;) Just figured I'd offer if it'd help.
>
> On Wed, Apr 18, 2012 at 20:10, James McDuffie<mcduffie at pitfall.org>  wrote:
>> On 04/18/2012 03:14 PM, Dan Buthusiem wrote:
>>>
>>> I can host it in a VM on my VM rig. Just let me know what you need.
>>
>>
>> Dan,
>>
>> When you say "VM rig" are you referring to something on a machine in your
>> house, or a dedicated machine at some data center which you have set up for
>> virtual machines?
>>
>> If its a machine at home I would err away from this solution since this is
>> exactly what Michael has currently running the list and something I think we
>> should move away from. Why? Well see slide 6 of my previously presented VPS
>> presentation:
>> https://docs.google.com/present/view?id=dfbp4hbc_24f679j7db
>>
>> I have an unused VPS sitting out there in cloud land that only costs me $15
>> a year. Theoretically we could top off my credits to keep that running for
>> years even if I were to be hit by a truck or something.
>>
>> What I think is best is that we set up some sort of redundant system where
>> multiple people can host backups of the main mailing list server using MX
>> records and rsync to help us out there. Due to the nature of the club I
>> think its best to come up with some sort of no single point failure
>> solution. I don't want the mailing list to go down because I didn't have
>> time or wasn't able to reboot a server or if some other bad event happened.
>>
>> Mailman looks non-trivial to set up and I am looking into what would be
>> necessary for a solution that allows seamless fall over to alternative
>> servers. Once I have that all understood I would build a shell script to
>> help "slave" nodes configure a mailman instance. But, it looks like this
>> will take a wee bit of time to figure out.