We're having a little difficulty with one of our servers.  I've dropped that server from rotation while I fix it, but some users might still get assigned to that server by their ISP's DNS cache.  That will resolve itself within four hours; I'm expecting to get the problem itself fixed sooner than that.

Update: Fixed now.  The server involved lost its NFS mount for some reason; it's been rebooted and put back into operation.

Posted by: Pixy Misa at 01:50 AM | Comments (5) | Add Comment
Post contains 76 words, total size 1 kb.

1 Hey Pixy, you seem to be having server issues again, I'm getting 500'd, its been spotty.

Posted by: DoublePlusUndead at June 24, 2007 10:27 PM (HgpT8)

2 Yes, I just got one myself.  I'm looking into it.

Posted by: Pixy Misa at June 25, 2007 12:03 AM (PiXy!)

3 I think I've found the problem.  We've been having network hiccups.  The software is supposed to handle this, but it isn't quite.

I should have a patch for this shortly.

Posted by: Pixy Misa at June 25, 2007 11:52 AM (PiXy!)

4 The patch for the intermittent database problems has been installed.  I ran it on the test system overnight, and it happily survives a complete reboot of the database server, so I think the problem is solved.

Of course, if the database server actually crashes, nothing will work, but momentary glitches should now be handled with rather more grace.

Posted by: Pixy Misa at June 26, 2007 07:53 AM (PiXy!)

5 There was another situation that the first patch didn't address, so I've applied a second patch.  All database queries now check the database connection and re-connect as required.  So this time we really shouldn't see any more of these errors.

Fingers crossed.

Posted by: Pixy Misa at June 26, 2007 12:38 PM (PiXy!)

Hide Comments | Add Comment