Tony

On Friday February 6th between 3:00 AM EST and 5:00AM EST we will be upgrading the firmware on the raid card on Skyline. We are doing this as a precaution has the firmware is a bit out of date and the problem experienced the day before should not happen. This portion of the maintenance will require about 15 minutes of down time as we cannot upgrade the firmware while the machine is active. The second portion of this maintenance window is to replace a bad drive the raid card had reported. Now it's no longer claiming to be bad but it was when the machine went down. So as a precaution we're going to replace the drive. This will not cause any down time and the array will rebuild itself in the morning. Date: 02/06/2009 Start time (EST): 3:00am End time (EST): 5:00am Duration: 2 hours Estimated Down Time: 15 minutes

Date: 2/11/2009 Start time (UTC): 06:00:00 End time (UTC): 11:00:00 Services affected: DAL 10G Verio Link Location: Dallas, TX Duration: 5 hours ================================================== Verio will be upgrading code on their router to increase network performance. Customer Impact: abnormal routing ==================================================

For quite a while the load problems have subsided and the array is rebuilding and there are no errors on sites or anything of that nature. This issue should be fully resolved at this point although we're going to investigate later today if we possibly have to replace anything that could have caused this issue to be on the safe side.

We already handle consulting for some company's but it is not something we advertise.

It seems the raid array is a little out of sync and is rebuilding. It's causing quite a bit of strain on the hard drives. We're looking into how to make this a little more tolerable.

The machine is online but as of this moment I do not know of the cause of the problem or if it should be fine going forward. I am awaiting an update from the on site technician who was working on the server.

Just an update we're currently attempting to bring the machine back online. The server was throwing errors about the raid array as well as the card itself not giving the best of messages. I should have another update within 15 minutes if it's not online as it will most likely require the card to be swapped out as it could be a possible cause.

As of this moment the Skyline server is currently off line and we are investigating the cause. We already have datacenter technicians looking into the problem and will hopefully have more information soon.

1. Assuming it's legal we have no issue. 2. It does not you would need to install it yourself.

Enough to handle the current workload.

Talking about it: http://www.hawkhost.com/About/past As for the name our old name had branding issues I saw someone selling the .com for hawkhost and bought it. Seems to paid off we doubled our subscriber base in less than a year and we should double again in 6 months or less based on projected client growth.

Are you making sure to use the right login? Keep in mind FTP logins created using cPanel have the form of [email protected]. The only login that does not have this is your main one. We cannot have everyone making real users otherwise we'd have hundreds of username conflicts.

We have a new site in the works so any forum changes will have to wait until then.

We say no more than 10% but it totally depends server to server.

newyear2009 for 50% off the first payment. s20percent for 20% off recurring.

Both locations have similar backbone providers and other specifications. The only real difference is they're about 30ms apart from each other. So users in Europe for example choose Washington DC over Dallas TX. For you I'd say there would not be much of a difference if at all between the two locations.

Nope we're not updating to the new site. My blunder I was testing something and left something important turned off thus breaking the site. I've fixed it.

As per our notice in October the old name server IP now has officially stopped serving dns requests.

cPanel automatically creates the dns records for the domain to make this possible. Keep in mind it's using a proxy to do so and it's not as secure as the https version of https://servername:2083/[url=https://servername:2087/]

503 errors are to due with the application not responding. As to the exact error I cannot say why it's not responding. It could be issues with PHP 5.2.8 which is recent as well or maybe LiteSpeed who knows. You're best to make a ticket because we can check logs for this sort of thing.

Unlikely to see Servlets any time soon as there are issues that we just have no figured out how to deal with. The big one being PHP and Perl run as the user the Servlets would not making it an absolute nightmare. It seems the best solution for Java is to look towards doing it yourself on a VPS.

Well we had some issues this morning that were not related to the occurring problem at random over the last few days. I've made some changes to the web server I believe I know why pages randomly became slow and I hope this change has now fixed it. I believe the priority at which the web server was running was not high enough. Being an event driven web server it's more sensitive to i/o tasks and when it does not have priority over the tasks causing it then issues start to come up. I have hopefully made so this will no longer be the case. What's more surprising is this has not affected every machine which is why it was not noticed earlier in deployment. We put it on machines CPU usage went down pages were served faster. Issue being each machine has different usage levels, patterns and the type of usage. A few machines were more CPU bound so they saw drastic improvement where load average dropped under 1 and speeds improved drastically. Others CPU was hurting them but it was not the only cause. We improved CPU usage and load was dropping quite a bit but i/o spikes were spiking load to reasonable levels but hurting the web server. So I believe it's fixed but I'm going to look for advice on more improvements. I believe our configuration is very unique compared to others who run this type of web server. We are not selling $25/month hosting accounts so obviously we're putting more users per server. So some of the benefits are negated as we're serving many more smaller sites compared to larger sites. But we do have some large sites who suddenly are not using near as much CPU which is great. Of course it still provides us drastic improvements in area's making it well worth it. I'm very sorry about the trouble the past few days but no ones perfect and we're not going to run at 100% uptime with no slow downs year round it's just not possible. We just try our best to solve issues and always look into any problem a customer may have.

Could not tell you without a ticket to look into the actual account. But one thing come to mind if you're receiving a lot of traffic you could flood out your PHP pool although unlikely. Or something else weird is going on. I know we did resolve something recently with regards to serving php pages.

I forgot to update this but this has been completed some time ago. Everything should be running as it did before except on a higher end machine and running LiteSpeed for the web server.

I'm going to take a guess and say you're on Mercury and I've already fixed the problem.