More downtime!

baddwin · Post by **baddwin** » Fri Mar 11, 2005 6:29 am

http://forums.worldofwarcraft.com/threa ... =1#post751

Hope they solve there issue...

Tacks · Post by **Tacks** » Fri Mar 11, 2005 6:37 am

Icecrown

LOLOLOLOL

Post by **Lyion** » Fri Mar 11, 2005 7:13 am

Frostwolf is my PVP server. Guess it's a sign I definitely need to not renew~

Tacks · Post by **Tacks** » Fri Mar 11, 2005 7:34 am

k c ya!

Post by **Lyion** » Fri Mar 11, 2005 7:44 am

Well, bye.

Tacks · Post by **Tacks** » Fri Mar 11, 2005 8:22 am

teehee, like I said before, I think they can afford a few (hundred) thousand players leaving. Wouldn't hurt my feelings in the least.

Post by **Lyion** » Fri Mar 11, 2005 8:26 am

Taxx wrote:teehee, like I said before, I think they can afford a few (hundred) thousand players leaving. Wouldn't hurt my feelings in the least.

You realize the people who leave are the ex EQ players, and the ones who stay are the Bnet ones.

I guess this is good from a PVP farming perspective.

I like WOW, but want to mess around with Jump to Lightspeed and the Planetside BFR Giant Robots.

I am on the only level when I have double exp plan for WOW. At this rate I should be 60 when the game is stable and since I rotate my levelling between Frostwolf and Whisperwind, usually one of them is up....

The Kizzy · Post by **The Kizzy** » Fri Mar 11, 2005 8:33 am

Or Medivh come play a paladin

Tacks · Post by **Tacks** » Fri Mar 11, 2005 8:41 am

Only ones left are the bnet crowd?

Hmm we seem to be doing fine in molten core...

Post by **Lyion** » Fri Mar 11, 2005 8:52 am

Taxx, you've way lost your sense of humor in regards to WOW. You take the game and any criticisms of it, real or imagined, way too seriously man.

Lighten up, Francis.

Tacks · Post by **Tacks** » Fri Mar 11, 2005 8:54 am

Do you understand this? I hope more people quit!

Tikker · Post by **Tikker** » Fri Mar 11, 2005 9:36 am

that's fine and all, but it's a fucking joke when 25% of their servers have been down for 3 days~

Langston · Post by **Langston** » Fri Mar 11, 2005 9:43 am

There obviously is a major problem with those servers. I hope they get them fixed. If you work in the IT field, you probably understand that sometimes servers crash... or that sometimes there's a programatic/code issue that causes massive problems. Not much you can do about it other than fix it, or sit around and whine about it. At least they're apparently trying to get it fixed.

I suspect those impacted will be given free time on their account. Blizzard has been good about giving that kind of recompense out after issues.

Tikker · Post by **Tikker** » Fri Mar 11, 2005 9:47 am

they've given 5 days for the affected servers

but I love that they initially described it as a minor network outtage

Post by **Lyion** » Fri Mar 11, 2005 9:50 am

Its really a simple premise.

Development -> QA -> Production.

There is really no excuse for extended downtime for a subscription online company anymore. Especially one that doesn't even host their own patching, but BitTorrents it.

Ugzugz wrote:If you work in the IT field, you probably understand that sometimes servers crash...

Occasionally, and they should be back up immediately. Everything should be redundant so there should be no downtime.

You know as well as I do this is a Blizzard fuck up. Its poor testing, rollout, and implementation.

Langston · Post by **Langston** » Fri Mar 11, 2005 9:57 am

The problem with that line of thinking, Lyion, is that only 25% of the servers are affected. If it was a major QA issue, I would think all of the servers would be affected since they all run essentially the same code.

I believe this is a database corruption problem. I don't know that for a fact... but it smacks of it. You're a DBA, what causes corruption on only *some* of a group of identical systems?

Tacks · Post by **Tacks** » Fri Mar 11, 2005 10:00 am

here's to assuming they're actually the same hardware.

I have a feeling they aren't since they pretty much scrambled to set servers up.

Langston · Post by **Langston** » Fri Mar 11, 2005 10:04 am

Well, honestly Taxx, the hardware is probably state-of-the-art and isn't the core issue. There are only so many types of processors, so many types of storage systems, and so many types of networks. I could see, though, that if they have a mis-configured or badly managed cluster environment that these issues could stem from hardware related problems. Otherwise, in all actuality, the hardware infrastructure makes little difference in the grand scheme of things - unless there was a failure. And if 25% of their hardware is failing, then the manufacturer that they purchased from is going to have a potential nasty lawsuit on their hands. Companies like this typically purchase Service Level Agreements for their hardware from the manufacturer where extended periods of downtime are actually billed back to that hardware provider. I know for a fact that EMC does this, as does Microsoft (not hardware, I know) for their Datacenter O/S implementations.

Tacks · Post by **Tacks** » Fri Mar 11, 2005 10:07 am

I'm saying in other words, the rate they set up new servers I think was way too fast to do any testing. I'll admit I know little about network infrastructures but I would assume something this complex isn't something you just throw up overnight.

Post by **Lyion** » Fri Mar 11, 2005 10:09 am

Ugzugz wrote:The problem with that line of thinking, Lyion, is that only 25% of the servers are affected. If it was a major QA issue, I would think all of the servers would be affected since they all run essentially the same code.

I believe this is a database corruption problem. I don't know that for a fact... but it smacks of it. You're a DBA, what causes corruption on only *some* of a group of identical systems?

Having a clueless DBA, or clueless SA set up the servers, or pushing QA servers into production. More likely the later. They probably share servers for these WOW instances. So when one database goes down the whole group of WOW servers is down.

They run Oracle, so if they mirrored their datafiles and had separate filesystems for their controlfiles they shouldnt have any issues at all, sans hardware failures. Having redudant hardware should help alleviate any hardware glitches to completely minimize downtime.

My guess is they fucked up the configuration of RAC for their Oracle servers, or they didnt setup their transaction processing good enough.

Also, due to their phenomenal growth, their standby servers may actually be running in production, without the proper memory to support them.

Post by **Martrae** » Fri Mar 11, 2005 10:14 am

They do seem to be running thru the Oracle people....isn't this the second one they've advertised in the last 3-4 months?

runamonk · Post by **runamonk** » Fri Mar 11, 2005 11:18 am

Lyion wrote:Taxx, you've way lost your sense of humor in regards to WOW. You take the game and any criticisms of it, real or imagined, way too seriously man.

Lighten up, Francis.

10 points for reference to Stripes.

chanteez · Post by **chanteez** » Fri Mar 11, 2005 12:33 pm

hyjal was one of the original servers, not one that was put up later to make room for the flux of people. So Ug's theory seems more realistic, since hyjal wasnt a quick fix server initially.

I wont whine about down time, I have had time credited in the first month of wow compared to eq did in 5 years or more i played with all their downtime. Plus Blizz is actually doing something about the problem.

Tikker · Post by **Tikker** » Fri Mar 11, 2005 12:35 pm

I don't ever remember EQ servers being down for the better part of a week at one time

runamonk · Post by **runamonk** » Fri Mar 11, 2005 12:44 pm

Tikker wrote:I don't ever remember EQ servers being down for the better part of a week at one time

Nameless was down for several days at one point, then they had to roll everyone back and rebuild it. SOE has had it's share of problems with servers, although I don't think they've had as many as blizzard lately.