STrRedWolf
Lazy-ass Drygerskunk
*pokes the server* Server kernel panic'ed again?
Takara_kitsune said:This is very interesting... very interesting indeed...
This is EXACTLY one week from the time that I posted that I was having issues with FA.
... Why do I feel there is a connection with the two events...? The time is EXACT!
Either this is a known high-traffic time that was intentionally hit...
Or perhaps there is a script that is set to trigger at this time that shouldn't be? You know, your standard overbearing CPU-eating maintenance routine. >.> Dunno if FA has a processor-munching maintenance program, but hey, it's always there as a possibility that there's a default one in place someone forgot to shut off... but this timing is TOO WEIRD.
Before anyone asks... yes, I do like coming up with conspiracy theories. >.> (As well as useful ideas at the same time. For every good idea, there are at least 3 not-useful ideas!)
It be a saturday night. FA gets hit hard with traffic at this time. I wouldnt be surprised if the damn thing blew up from the load.
Charem said:This Saturday seems to have been overly busy in numerous ways...Safeway, where I work, was nearly brought to its knees with all the people that showed up to shop yesterday, and the same problem happened with my roommate's restaurant job...so I'm not surprised if FA was similarly overwhelmed.![]()
Fara said:Too many people in the world, maybe? I blame the artists for being so excellent, and attracting so many people in. Of course, if the downtime really bothers you, and you'd like to see FA being more reliable, there's a very simple way.
Donate.
STrRedWolf said:*grumbles* Can't we get a second web server to handle the load and fail-over for FA?
Petrock said:Too many people on at the same time = accidental DDoS I guess. XDD
Neh. I'd say get a mirror set up or something to absorb some of the workload. Maybe have the backup server take some load on saturdays, then go back to just being backup on other days?
guyver47 said:...
As long as my notes and everything in the message center remains as it was before the downtime, I'm satisfied.
...
tsawolf said:guyver47 said:...
As long as my notes and everything in the message center remains as it was before the downtime, I'm satisfied.
...
Theoretically speaking, the only data that could possibly be lost would be artwork.
Notes, everything like that, it's all stored on the database server, which is still chugging along happily (much more happily than usual, since it has a total of 0 requests to fill).
I am not saying that there was data loss. Just that it wouldn't even be possible to have that kind of data loss in this scenario.
tsawolf said:To be honest, in the log samples that we took a few weeks ago, there wasn't an appreciable increase on the weekend. Load was pretty much constant across the week.
Another server could certainly be setup as a master/slave webserver, but we'd need some way to keep the images constant across the two servers. That would mean some tricky footing, but certainly wouldn't be impossible.
Only problem: Servers are expensive. Very expensive. And if you want to do things right, you can't skimp on the money. Part of our problems are generated by not using the best hardware out there (this particular downtime, I don't think is related).
If someone has a few spare thousand dollars they want to donate... :twisted: