[TriLUG] TriLUG Server Refresh

[Craig Cook via TriLUG]

>Not sure what access to the Datcenter is like for physically replacing, we can prolly assume 1 week.


This suggests to me you probably want a RAID 5 setup with at least one extra disk standby.  If you plan to run 1 week in degraded mode there is a reasonable chance you will loose a second disk.  You also need to consider the RAID rebuild time.  If it takes 24+hrs to bring a new disk into the mix, you need to allow for that as well.

So, this can be called RAID6, or RAID5 with a hot spare.  Two slightly different things.

If you are going to have 2 redundant disks then you will need to purchase 5 disks.

If you don't know who is going to support the OS then you probably want a "popular" one, like CentOS or Ubuntu.  Ubuntu tends to have later versions of packages available is a little easier to manage than Redhat based systems.


Craig

________________________________
From: Jonathan Mainguy <jon at jmainguy.com>
To: Craig Cook <cncook001 at yahoo.com>; Triangle Linux Users Group General Discussion <trilug at trilug.org> 
Cc: "trilug-request at trilug.org" <trilug-request at trilug.org> 
Sent: Friday, May 1, 2015 12:58 PM
Subject: Re: [TriLUG] TriLUG Server Refresh



Pilot takes up about 150GB right now. So 1-2tb should be good

Not sure what access to the Datcenter is like for physically replacing, we can prolly assume 1 week.

SC - Can yall comment on budget for disks?

At the moment we only plan on running Pilot, but of course want to add more VM's as we have time to move services off, guestimate around 10?

Support - SC changes every year, and volunteers come and go, so we can figure out who will support it for the near future, its hard to know who will support it 12months out. That being said, we should choose something we can train new volunteers up on easily.




On Fri, May 1, 2015 at 12:49 PM, Craig Cook via TriLUG <trilug at trilug.org> wrote:

Before you jump into buying disks, how about some more questions?
>
>How much disk space is needed?
>When one hard drive fails, how long do you expect before someone can physically replace it? (The longer the delay, the more built in redundancy you should plan for).
>
>What budget is available to purchase disks?
>
>How many VM's do you plan to run and what do they do?
>Who is going to support the OS/Hypervisor? (What skills to they have?)
>
>
>After questions like these are answered you will get a better understanding of what I/O profile is needed.  ie. running a heavy duty I/O bound DB server may need SSD's.  Running a web server serving a few hundred people a day probably doesn't need SSD's.
>
>Craig
>
>--
>This message was sent to: Jon Mainguy <jon at jmainguy.com>
>To unsubscribe, send a blank message to trilug-leave at trilug.org from that address.
>TriLUG mailing list : http://www.trilug.org/mailman/listinfo/trilug
>Unsubscribe or edit options on the web  : http://www.trilug.org/mailman/options/trilug/jon%40jmainguy.com
>Welcome to TriLUG: http://trilug.org/welcome