Tony

Go with Dallas it has a new machine deployed that is twice as fast as anything else we have right now (extra CPU, more ram and faster hard drives) Washington DC has yet to see this new configuration yet.

The machine is back online and booting up all services. Here's hoping that's the end of it.

All the fans were working fine but we replaced the raid card anyways as it was not running at an optimal temperature which could suggest a problem. We are now running a fsck which could take up to an hour to do.

So much for doing this in a decent timeframe. We're going to take it offline right now as it's just gone and blown up again. We'll also be running a fsck on the partitions after the raid card is replaced as we have no idea what damage it could have done. The raid card replacement will be 30 minutes the fsck will probably take an hour to do. Unfortunately this is necessary to once again have a stable system.

The machine will be going offline at about 11:00 PM EST and will be down for hopefully no more than 30 minutes as we'll be checking the fans on the machine to make sure they're all running. If any are not they will be replaced at that time. As well we'll be replacing the raid card as a precaution as even if it's good we'd rather not take the risk.

The machine is online at the moment it would be appreciated if everyone does not go generating backups as they will fail and cause us more problems as we prepare to take the machine offline to investigate and solve the root cause of these issues.

After dealing with several instances of file systems on the neptune server going read only we're doing emergency maintenance to hopefully solve this issue. The raid card on the system is running abnormally hot so we believe there could potentially be problems with cooling on the system or the card being bad. We will be taking the machine offline within the next hour to get to the bottom of this issue. We'll be updating this post as we have more information such as when it goes down and the amount of time it will be down for.

It's tough to do when you are coordinating with a artist for drawings another person for the site design then finally having programming done for it and content created. There were delays on numerous stages resulting in delays for the others. We have the programming side basically done and now it's just a matter of cleaning up the content.

There is something I forgot to note while doing the replacement we did notice the temperature of the raid card was a little higher than normal. If the drive replacement does not solve the issue then we may explore taking the machine down for up to 30 minutes to investigate if there is a heat issue. It is unlikely we'll do this as we believe it's the drive but it's just another possible cause.

The drive has been replaced and the raid array is rebuilding.

The neptune server is currently showing one of the drives in it's raid array as having bad sectors. This is causing file systems to go into read only mode as the system is unable. The replacement of the drive should cause no down time what so ever but we've already went and rebooted the machine as we'd like a clean slate for the drive replacement. We'll update the topic as there is news about the replacement.

For anyones curiosity with wordpress it'll in apache take over on the hotlink protection when you use mod_rewrite. To get around this you'd need to put "RewriteOptions inherit" which will fix it. It was introduced in the latest lsws release we're using.

Was back up and running in under 10 minutes

We're rebooting the Neptune machine right now to resolve a problem. It should be back up in about 10 minutes. We'll update this topic when it is back up.

It's a bug in LiteSpeed with handling multiple levels of rewrite rules. It'll be resolved before the final 4.0 release (we're running RC1 right now).

Have you made a ticket about your issue? Hotlink protection should work fine so we'd like to investigate this but you'll need to make a ticket.

This work has been completed.

On shared web hosting you're going to need to make a ticket so that we may add the records.

It's relatively easy to do. It's not as easy as say if the previous host was using cPanel but still easy. With the other host being cPanel based we can actually restore everything. With a host like ipower it's best the site owner or someone with access to it does it as the formats of usernames for the database and possibily even the structure of the account will be different. So I guess just a pain to do manually but once you're moved then no headaches

What you'll need in order to move the files is ftp across all the PHP files and images. So the best way to do this is to FTP into the space with ipower and download all the files in the public_html portion. Then you'd login to your ftp with us and upload the files to the public_html folder. You'll need the database backups which you have. You'll create the databases for each script. You'll then import them into the database. Finally you'll need to modify the configuration files of the scripts to point to the new database names. After this in theory everything should work.

Drive has been replaced and is rebuilding. It'll take a few hours but this should not be noticable.

Within the next thirty minutes we will be replacing a bad drive on the Jupiter server. We are doing this as a precaution to problems with this drive yesterday which resulted in us having to reboot the machine and then a bit of work for the drive to be found again. The raid card is currently reporting the drive as healthy but we're not taking any chances. There will be no down time as this server supports the swapping of hard drives while the system is live. Once the drive is replaced we will rebuild the raid array which may cause a higher load than normal on the server but nothing to drastically slow down web sites. We are very sorry about the lack of notice on this maintenance window. We acted yesterday on the problem and the drive came up fine today. But we're playing it safe in case the drive is faulty it's just not showing up consistently. We'd rather have it replaced now then deal with data corruption further down the line or more problems. Date: 02/10/2009 Start time (EST): 2:00pm End time (EST): 4:00pm Duration: 2 hours Estimated Down Time: None

I think the best measure of any account is page views and not online users or anything of that nature. As for the # it's going to vary server to server as some there is a lot more room for users to use resources. But a general rule 50,000 page views would be a reasonable amount. But it also depends on what you run we've had people hosting forums with 75 users online and we had no idea. The the inverse we had users with just 5 online at any time and they were using half the server resources with a shoutbox. So basically a forum starting out I doubt you have much to worry about and we'd contact you when we think it's getting to the higher amount of CPU and page views. The big thing about making so you can handle more traffic is not having silly add-ons like shoutboxes which will increase pageviews from say a 1000 a day to 20,000 just due to the fact there is constant refreshing.

Even with the up to date information it bases things off of the information in the IP database SWIP I believe? So for example our Washington DC location claims to be in Dallas because SoftLayer's corporate office is in Dallas and they own the IP's. If we put our information on the IP's then the servers would be in Canada supposedly so we leave Softlayers information.

The raid array has been rebuilt and is optimal once again.