
Ouch! This started as -500 errors thanks to poor web host planning, and thanks to some poor work by our host Jaguar PC it escalated into total site failure over 2 days.
Anyway, we ended up changing to a ServInt.com VPS account, and our problems look like they're all fixed now. I threw the details of this episode into the comments; perhaps others will find that valuable in some way.








Testing... in response to advisory involving frequent -500 errors when editing posts, MT scripts ending prematurely, comments posting late.
I think we've found the problem - all MT users take note. A wayward cPanel update was killing us. Fixed thanks to this
I hate computers.
Will, heh :-)
As noted in the entry above, unfortunately cPanel is pretty determined to do the upgrade of DBD::mysql, thus re-creating the problem.
We've created a script that downgrades it automatically, and added it to our crontab so it runs every 2 hours (it's not a very resource-hungry script). Meanwhile, we're hoping Six Apart will have a fix for this soon so that cPanel can upgrade the DBD::mysql library without breaking MT.
I mention this here in the comments because we may re-use this post for future technical updates etc. and I want to preserve the information re: this widespread problem.
0001001110010110001100101011101001010
0010001010110100110010110100110010101
1010001001011101001010010101001011001
1010010100100100010100101101001001110
Our technical difficulties with -500 errors (back end) and slow server response time (reader end) returned, albeit for different reasons this time. The basic problem is that JaguarPC's server didsn't have enough RAM, so the VPS accounts have no overhead capacity and choke when they hit their guaranteed minimum limits.
Virtuozzo told me we were regularly redlining system resources, and I got lots of "Black Zone" QoS alerts. When we go Black Zone, we get -500 errors on the back end and comments may fail. Yellow and Black Zoning CPU and System resources will also make the main HTTP site slow.
For the really geeky, note that most of the Black Zone problems involved "privvmpages" (private memory pages) errors, involving the size of private or potentially private memory allocated by an application. The memory that is always shared among different applications is not included in this resource parameter. There's also the occasional "numtcpsock" (number of TCP sockets) and "dgramrcvbuf" (total size of UDP and other datagram receive buffers) error.
We ended up shifting our account within JaguarPC, which they completely botched so our site was either down or inoperative all day Friday.
That prompted a move to ServInt as out new hosts, and they've been quite helpful. So far, so good.
So what did you say or do to get them so rattled that they completely self-destructed and all their servers are currently down? Either they are fixing it and it will be better than ever (there is no question it is cheap) or they're in trouble and folks may need to make plans accordingly. -- It is my opinion that Greg will get it fixed.
Hank Roth
(If anyone hears anything about the problem at Jaguarpc getting fixed and when, write me)
Not trying to be picky or anything but your new hosts say they allow root logins, which as I understand it is not such a good idea to allow root to login from remote.
Hello,
We started with Aletia Hosting which was bought out by JaguarPC. During thid time, we have experienced a number of problems. Sometimes they are resolved quickly (as so they should be), in other situations it will take hours or days to resolve.
We have left JaguarPC and returned. Now we believe that will leave for good. I am currntly looking at a blank screen where our site should be. I even tried to check JaguarPC's site and it is gone as well.
We pay a large amount of money to ensure this type of problems don't occur.
If you are thinking of JaguarPC as your hosting service, I would reconsider. Do some homework and find someone more reliable.
Regards,
Jeff