Server move next week; brief outage

There will be a brief outage late Monday night/early Tuesday morning as my server is moved to a new datacenter. The outage is expected to last about 20 minutes and will take place some time after 10 PM PST (1 AM EST). If all goes well, you’ll barely notice.

FRN down due to MySQL outage

Once again, DreamHost MySQL server amber is experiencing an outage, which means, unfortunately, so is FRN. This is the fourth major outage of amber in the last month (1, 2, 3); I’ve asked for an explanation.

Update, 1:25 PM: Seem to be back now; downtime was about an hour and twenty minutes.

Yet another database outage

MySQL server amber is down again; see previous entries for impact and background (1, 2).

Update, 1:40 PM: The sites started coming back a while ago, and DreamHost signalled the all clear about 10 minutes ago. Total reported downtime just over two hours, though actual site inaccessibility was less than that.

More issues with amber

As before, sites using “amber” as their MySQL server — notably, FRN and Ontario Herpers — are unreachable at the moment: “amber” does not seem to be down, but the pages are simply hanging forever when trying to load them.

Update, 10:43 AM: The outage is reported resolved, and FRN (at least) is back online. Total downtime, about an hour.

MySQL server issues

MySQL issues are affecting several of my blogs this morning. Blogs published with WordPress or Textpattern, which are published dynamically, may be unreachable from time to time: they’re getting sporadic connection errors. Blogs published with Movable Type, such as The Map Room and The McWetlog, will continue be reachable, but posting, editing and commenting appear to be touch and go at the moment, or at least really slow. Blogs published with Blogger should be unaffected.

From what I can tell, they’re on the job. With any luck this will be resolved shortly.

Update, 10:45 AM:

Two hours after the first outage was reported (not by me), MySQL server “amber” is completely down.

This means that Reptile Breeders Group (or at least those portions not currently being autoforwarded to my personal site’s Reptiles section) and Ontario Herpers, as well as a new project I’m not ready to announce, are completely unavailable.

The database powering Reptiles at Risk is also on “amber,” but the site is not dynamic and is still up (though not changeable).

Most of my other Movable Type-powered blogs use a MySQL database hosted on another server, which is a bit slow right now but still responding.

12:00 PM: Crap, amber is still down, and has been for three hours now. The other database server is working much better, though.

12:12 PM: Don’t look now, but we appear to be back. For now, at least.

Sites down; changing server on Sunday

The server on which my sites are hosted is reporting abnormally high load averages, which is what usually happens when my sites are unavailable, which they have been for the past half hour.

During DreamHost’s network problems, they moved me to my current server a month ago. Since then, I have been experiencing short outages associated with load-average spikes: the outages last about 10-20 minutes, and occur between one and three times a day. (The current outage has lasted longer and appears to be much more severe.)

It turns out that I should not have been moved to that server, which seems to be even less stable than my previous server. So we’re moving again — on Sunday afternoon (2 PM PDT, 5 PM EDT). The move should take an hour or so to complete, and there may be some brief downtime associated with it. Hopefully, though, it’ll put an end to the periodic outages I’ve been facing for the past month: they’ve been short enough not to cause serious problems with my sites, but frequent enough to be annoying.

As for the current outage, it seems to be abating at the moment (8:55 AM EDT).

Update, Oct. 1 at 11:40 PM EDT: The server move did not take place as scheduled. DreamHost apparently had some trouble today (I was out and about so missed the opportunity to freak out about it) and I suspect I got lost in the shuffle. I’m still raising hell about it.

Oct. 1 at 11:48 PM EDT: No sooner had I written the above … DreamHost has initiated the server move. It should be on a new server within the hour, if all goes well.

Scheduled downtime redux

DreamHost is moving filers tonight between 1 AM EDT (10 PM PDT) and 4 AM EDT (1 AM PDT). My sites may be down for a brief period (they say around 30 minutes) during that time.

Scheduled downtime

DreamHost has announced that additional downtime for maintenance is scheduled for tonight at 1 AM EDT (10 PM PDT). I anticipate that my sites will be unavailable for a short while around that time.

Another major outage

My sites have been down for the last 15 minutes; DreamHost’s web panel is also down, so it’s probably a significant outage this time. Sigh.

Update, 11:40 AM: Everything is now back online. That one took everything on DreamHost offline for a bit.

11:45 AM: Back on the roller-coaster. My sites are down again.

11:50 AM: Up again.

12:55 PM: Everything continues to work well. DreamHost reports that a router “hiccup” was to blame.

Network upgrades unsuccessful; sites down again

DreamHost reports that their router upgrades weren’t as successful as they’d hoped, and indeed problems continue: my sites have been down for the last 10 minutes or so (since around 7:20 PM). I’ve had periodic outages for the past week or so of a few minutes’ duration; hopefully it’s one of these, rather than a clusterfuck like the one we faced on Friday night.

Update, 7:40 PM: Back after an outage of 20 minutes or so. Hopefully it’ll stay up for a while.