Our host has 4 levels of VPS hosting
Tier/CPU/Memory
Tier 0 - 1 CPU - 750mb
Tier 1 - 1 CPU - 1750mb
Tier 2 - 2 CPU - 3500mb
Tier 3 - 4 CPU - 7000mb
The night of the NC State game, we decided to upgrade to the next level of hosting(Tier 1.) We have been on Tier 0 since we moved to our new host in March 2012.
The slowness started with the first migration and has been occurring ever since. I've been talking with support for a few hours a day since Tuesday. They first suggested that we move to the next level(Tier 3), because we had uncovered a CPU bottleneck when our memory increased. This was plausible, because we were having memory issues on Tier 0. We moved to the next level and it didn't alleviate our problems. Tonight, during the GT game, our site was performing like dog shit, and people were complaining. I quick re-sized our server to the next level of hosting(Tier 3) During this process, server load and memory usage dropped to pre-upgrade(normal) levels. Very intriguing because the Tier 3 CPUs and memory hadn't been engaged at this point.
I, of course, inform our hosts of this chance occurrence. I had been yelling at them all week because the only thing that had changed on our site was the fact that we upgraded our server. Of course, they told me that it wasn't their fault. Well, as soon as things returned to normal during our upgrade tonight, I emailed them to look into it.
They have since responded that we shouldn't change anything because load/memory usage is perfect. Tier 3 hosting for OGBoards.com is like killing a mouse with an atomic bomb. It isn't necessary. We would need 2,000 more active users to need this type of hosting. In addition, it would cost $300/month. That type of monthly expenditure isn't responsible or sustainable.
Here are some of the graphs from our backend.
http://i49.tinypic.com/344bm2u.png - Load Average 2 weeks
http://i48.tinypic.com/sfkt8y.png - Bandwidth Average Month
http://i46.tinypic.com/5v3g7.png - Ram Usage 2 Weeks
http://i46.tinypic.com/sfdz4n.png - Active Process 2 weeks
http://i47.tinypic.com/11ltr1z.png - ACtive Process 2 weeks
Basically, these graphs show that the issues were on our hosts side of the fence and that they have been dicking me around all week. The steady pre-spike values are all Tier 0 hosting.
It's 1am on Sunday morning and I've started the reversion to the Tier 1 server. Ideally, we will pick up where we left off tonight, but if there is any indication that performance is suffering, we will revert back to Tier 3 and continue working with our hosts to figure out the problem.
I hope this has cleared some things up for you all. We sincerely apologize for the service interruptions over the past few days.
Tier/CPU/Memory
Tier 0 - 1 CPU - 750mb
Tier 1 - 1 CPU - 1750mb
Tier 2 - 2 CPU - 3500mb
Tier 3 - 4 CPU - 7000mb
The night of the NC State game, we decided to upgrade to the next level of hosting(Tier 1.) We have been on Tier 0 since we moved to our new host in March 2012.
The slowness started with the first migration and has been occurring ever since. I've been talking with support for a few hours a day since Tuesday. They first suggested that we move to the next level(Tier 3), because we had uncovered a CPU bottleneck when our memory increased. This was plausible, because we were having memory issues on Tier 0. We moved to the next level and it didn't alleviate our problems. Tonight, during the GT game, our site was performing like dog shit, and people were complaining. I quick re-sized our server to the next level of hosting(Tier 3) During this process, server load and memory usage dropped to pre-upgrade(normal) levels. Very intriguing because the Tier 3 CPUs and memory hadn't been engaged at this point.
I, of course, inform our hosts of this chance occurrence. I had been yelling at them all week because the only thing that had changed on our site was the fact that we upgraded our server. Of course, they told me that it wasn't their fault. Well, as soon as things returned to normal during our upgrade tonight, I emailed them to look into it.
They have since responded that we shouldn't change anything because load/memory usage is perfect. Tier 3 hosting for OGBoards.com is like killing a mouse with an atomic bomb. It isn't necessary. We would need 2,000 more active users to need this type of hosting. In addition, it would cost $300/month. That type of monthly expenditure isn't responsible or sustainable.
Here are some of the graphs from our backend.
http://i49.tinypic.com/344bm2u.png - Load Average 2 weeks
http://i48.tinypic.com/sfkt8y.png - Bandwidth Average Month
http://i46.tinypic.com/5v3g7.png - Ram Usage 2 Weeks
http://i46.tinypic.com/sfdz4n.png - Active Process 2 weeks
http://i47.tinypic.com/11ltr1z.png - ACtive Process 2 weeks
Basically, these graphs show that the issues were on our hosts side of the fence and that they have been dicking me around all week. The steady pre-spike values are all Tier 0 hosting.
It's 1am on Sunday morning and I've started the reversion to the Tier 1 server. Ideally, we will pick up where we left off tonight, but if there is any indication that performance is suffering, we will revert back to Tier 3 and continue working with our hosts to figure out the problem.
I hope this has cleared some things up for you all. We sincerely apologize for the service interruptions over the past few days.