?

Log in

No account? Create an account

Previous Entry | Next Entry

up. late.

Working on bug fixes and fix verification for a site for centric. I've been on it for almost four hours, now. Good progress. Got pretty much everything checked off, and figured out one "hidden" bug (that I'll probably forget, so I'll mark it here--dev server and live server seem to have different magicquotes settings or something; they need a different number of stripslashes! WTF!? Eep.)

Meanwhile, poemranker has been crashing on and off. It's really obnoxious. I don't know what's up. Key word: org.postgresql.util.PSQLException: Unknown Response Type 9 -- "9" changes every page reload. Searching for this brought up:

http://archives.postgresql.org/pgsql-admin/2005-08/msg00237.php

I'm so tired I didn't realize it was me until I read my sig. Arrrgh! Yeah, I'd thought nobody had replied to it. I was right. =/

On a sidenote, it would be spiff if google could guess when a search result was an archived thread, and sort by date. I mean, _I_ can tell, so why can't it? ;)

God damn! Searching for this result actually turns up poemranker pages at the top of the list. Now is that wrong, or is that just WRONG!?

Ooh, this thread just before me says it could be a response to an out of memory error. That would be interesting. Top, of course, doesn't show anything out of memory. It appears I'm only giving java 256M, though. But... setting -Xms and -Xmx to 512M doesn't seem to change usage in top. =/

Comments

( 1 comment — Leave a comment )
zotsf
Sep. 30th, 2005 04:33 am (UTC)
little bugs
heh, we just discovered a fix to a bug that has been plaguing us for 6 months now.

Our production servers were tossing IO errors and breaking connections to db cluster, thereby tossing one db out of sync, which results in it being kicked unceremoniously out of the valid cluster. Oi, then resyncing needs to take place and developers get grumpy.

My architect kept swearing that it was a network issue...bad network..bad network...was the mantra.

I have learned that if my architect swears it is a sys admins problem, then it will more often than not turn out to be his problem.

Turns out I was right. There was a bug in the jdbc driver we were using that was the cause of this. A recent release has fixed the issue.

heh...little bugs.
( 1 comment — Leave a comment )

Latest Month

February 2016
S M T W T F S
 123456
78910111213
14151617181920
21222324252627
2829     

Tags

Page Summary

Powered by LiveJournal.com
Designed by chasethestars