Switch Update (Selma server issues)

The issues that arose yesterday haven’t surfaced again. The affected servers were venkman, cerritos, archibald, flanders, rupee, ralphie, cheeky, pizarro, butterfinger, and selma and they all seem to be doing much better with the exception of selma.

Selma was moved to an entirely different rack last night to see if we could isolate the problem. It crashed a couple more times and this morning we discovered an IP that was crippling selma with an attack of some sort. We’ve since firewalled that IP and selma seems to be doing much better, we’re keeping a close eye on it (as well as the previously affected switch) to make sure these troubles are behind us.

We appreciate your continued patience through these strange issues, if I haven’t specifically mentioned your server above then any problems you are having are unrelated and you should contact our tech support team.

8 Responses to “Switch Update (Selma server issues)”

1. 1
   Frugal Says:
   October 6th, 2006 at 2:18 pm

   My site is still DOWN, and has been down for 22 hours.

   Is it so hard to ask for a restore of my mysql database?
   My problem is probably not related to the switch, since I’m on aviation.

   Thanks.

2. 2
   anonymous Says:
   October 6th, 2006 at 3:42 pm

   The problems with selma, selma it seams to have some problems after the harware switch.
   They mess it up. It was working until they make the changes, i suspect thet it has some think to do with the new configuration, the hard and the soft dose not stick togeder very well.

3. 3
   Nicholas Lee Says:
   October 6th, 2006 at 4:41 pm

   Seems like DH has had one of those weeks which can make or break. Critical system failure is pretty nast, especially if there a DOS going on as well.

   However, not segmenting your internal and external networks seems like a bad call. When you load is high and one part of the network start flapping, then your whole system goes down. At least if your disk io traffic is on an internal backbone you can segment servers into different pools and something like an attack or switch failure isnt going to take down everything.

   I just signed up on the 3rd, so I’m not too worried at the moment. Plus a good shake out now is probably what DH needs to get there system structure working again.

4. 4
   brasscrest Says:
   October 6th, 2006 at 5:14 pm

   EVERYTHING did NOT go down. They have over 1000 servers. About 10 were having issues related to this particular issue. ONE was targeted by the DoS attack. One-tenth of one percent is not everything! Even one percent is not everything.

5. 5
   Jared Says:
   October 7th, 2006 at 10:36 am

   Dreamhost is doing a great job, I believe. Just be patient. They are working very hard on fixing the issue. If they don’t reply to your ticket, then they already know about it and are trying to fix it at the time. The outages are automated, they singal tech support when they confirm the site services are down. Otherwise, it will wait an hour and if nothing happens, it declares it OK.

   If you want to have it fixed ASAP, do us a favor and apply for a job. You probaley won’t know how to fix it anyways, and it would take you days to fix it, rather than hours like DH.

   Don’t blame Dreamhost for it. They are trying their best, but it seems like the whole internet is starting to fall apart. Plus, all this is due to the effects we have commited on our antosphere, causing blackouts and network failures.

6. 6
   Jared Says:
   October 7th, 2006 at 10:37 am

   If you want to restore your DB, go to the MySQL managerment and click restore. xDD

7. 7
   mom Says:
   October 8th, 2006 at 12:09 am

   فتح افلام وصور

8. 8
   Blaze Says:
   October 15th, 2006 at 6:10 pm

   How many servers are there? Geeze.