Brett Parker wrote:
On 06 Sep 10:25, Adam Bower wrote:
On Mon, Sep 06, 2010 at 10:00:34AM +0100, MJ Ray wrote:
includes IMAP. http://www.software.coop/products/web/#hosting We monitor this and it shows 99.147% availability this year so far.
That's just over 2 days downtime, hopefully it wasn't all in one go. ;)
No, the above figure is email service *availability* (which seemed most relevant here) and not the host *uptime* which most sites state publicly - that's 99.96% and the 2h downtime so far was mostly due to a disk changeover and testing.
Of course, that's a totally useless statistic, it does not, for example, state wether the unavailability was due to scheduled maintainence etc.
I'd have thought better from a stats postgrad... ;)
Why? Stats postgrads think people on the R email lists should be able to cope with the unvarnished statistics. ;-)
But seeing as someone's interested: I'm pretty sure the lost service time was all due to a hard disk failure that slowed the RAID array to the point where email service was intermittently unusuable (we use Maildir, which I think is more disk-intensive but lighter on memory), followed by scheduled weekend downtime to fit a new disk. There was a complication and finding and removing the disk that was timing out from RAID wasn't straightforward.
Several high-priced *guaranteed* services I've seen only payout under 99%, so I don't think the co-op's *standard* hosting's email availability being over 99% is too bad. We've done better before and we'd've liked it to be better this year, but hardware failures occur at random. Following the above-mentioned problem, we've decided to replace other hardware in the near future and get it all back on kit that's under warranty. That runs counter to my usual "if it ain't broke..." attitude, but experience shows more recent servers are usually down less.
Does anyone have suggestions for other actions that would push that uptime higher while keeping the cost reasonable?
Thanks,