Sphinx

[stratogod]

I recently switched hosting services to you guys, but as I have been moving backups, files, databases over I have had several issues with Sphinx just flat out dying...? meaning I can't connect, view, or do anything. This has been going on for a couple of days.

I'm on a semi-dedicated account, and the whole idea of moving to JaguarPC was to alleviate all the past server problems. Is this a typical problem or are you guys doing some maintenence? Is it possible that the large backups I'm restoring is causing the server to cough? I mean sometimes it is down for more than an hour. Please help, much appreciated.

http://sphinx.nocdirect.com/

[stratogod]

Ok --- it seems to be spitting like fire now... and everything is fine. I just hope this type of thing isn't regular because I thought on SD accounts that it was restrained to less thatn 5% of the amount of users on a shared environment.

Just curious on why over the last 2 days I have been having spurts of inactivity and unresponse from the server...

as I said, it is working great now.

[jason]

As a five-year JPC veteran I can say that what you are experiencing is not typical. Most of the complaints about poor performance on normal shared machines comes from brand new clients. JPC puts new accounts on new servers, so I suspect some of performance problems come from having a whole bunch of people trying to move a whole bunch of "stuff" onto the servers at once. Once they stop adding clients to that machine the complaints seem to decrease.

I work for a university and I equate this to Freshman Move-In Weekend. Normally we have a pretty quit campus. Students come and go at will, but we have pleanty of roads and wide walking paths to accommodate them, so getting around isn't a problem. However, there is one weekend every year when that all changes. Suddenly, for about three days in the middle of August we have hundreds of people who have never been here for more than a few hours converge on those roads and walkways. No one knows where they are going, so they move slowly as they try to read signs and maps. Not only that, but for every student there are usually between 2 and 6 more people with them--parents, siblings, even grandparents in some cases, and all of those people are moving suitcases, boxes, computers and furniture around. Then, as quickly as it starts it all stops and life goes back to normal.

I wouldn't expect a SD plan to have quite the same stress put on it that a shared server would, but if enough people are moving enough data around, I suppose it could happen. If you continue to see issues I'd recommend opening a support ticket so that support can check to see if they can better optimize things.

--Jason

[piper]

I just got a semi-dedicated account on Sphinx and I'm having the same problems. At the moment it is dead. Just before I checked my CP and the server load was 118!

Exact Stats:
Server Load 118.48 (2 cpus)
Memory Used 97 %
Swap Used 100.00 %
Disk hda6 (/) 53 %
Disk hda1 (/boot) 20 %
Disk hda8 (/home) 38 %
Disk hda5 (/tmp) 14 %
Disk hda2 (/usr) 31 %
Disk hda3 (/var) 38 %

[wgiese]

Sphinx has been acting up for weeks - I've opened numerous tickets. My users have notified me almost 50 times in the last month regarding the outages. Prior to a month ago, it was rock-solid, so something has changed.

Typically I'll give support a chance to work things out, but the problems have gone on too long, and one expects a little more when paying for SDX. I've PM'ed Greg...hopefully somebody will look into things a little more deeply. To date, all I'm getting is the "it's fixed now, try again" response from support.

It's down again as I type - 2nd time I've noticed it today.

[wgiese]

Just before I checked my CP and the server load was 118!

It's not just server load - MySQL has failed on me twice in the last 10 days or so due to /tmp being full.

[piper]

Oh man... Just when I thought my website troubles were over with this new semi-dedicated plan...

[JPC-Greg]

We'll get this sorted out and develop a plan to catch this sort of thing before its gone this far.

[piper]

Support has apologized for the inconvenience and has told me Jag is building a new SDX machine that I can get moved to once it's finished.


* Man, I always get the dying servers, lol.