Too many accounts on the server, 543 to be exact

[dhoom]

When I joined the number of accounts per server were approximately 140, today the number is 543 to be exact. Free hosting service providers like freehostia etc. have approximately 690 per server.

Just a concern about the quality of service.

Thanks,

[Karl]

The number of domains does not accurately reflect what the proper load on a system is. We have many customers who hosts dozens, if not more, of domains on a single account, often pointing to the same content, etc. Not to mention that web5 is a very high end system with 4 processing cores, 8GB of RAM, and drives ina RAID 10 array. A much more accurate way of seeing if a server is overloaded is checking the resource usage, which you can do yourself via SSH.

If you'll look at the stats of web5 you'll see the average load is less than 2, and on a 4 CPU system such as that server is that would be considered low load. There is also nearly always 1GB of free RAM and disk IO is kept at a minimum. Looking purely at the resource usage, the system is not even close to being overloaded.

As a note, you may have experienced issues over this past week. We had announced a bug with Litespeed last week and that issue has since been resolved. Things are currently operating smoothly and should continue to run smoothly.

[NickCatal]

Yeah, if you look at the load of Web 5 it is pretty small.

It isn't like every one of those domains is getting a massive amount of traffic either. I have like 7 domain name 'accounts' and only really 'use' 3 of them on any regular basis. The rest are just there, but nothing useful is on them.

[larry]

Yeah, I agree it isn't overloaded either... in fact the load average right now is only 1.3 which isnt very much :D

I checked last night and my site loaded quickly

[dhoom]

The number of domains does not accurately reflect what the proper load on a system is. We have many customers who hosts dozens, if not more, of domains on a single account, often pointing to the same content, etc. Not to mention that web5 is a very high end system with 4 processing cores, 8GB of RAM, and drives ina RAID 10 array. A much more accurate way of seeing if a server is overloaded is checking the resource usage, which you can do yourself via SSH.

If you'll look at the stats of web5 you'll see the average load is less than 2, and on a 4 CPU system such as that server is that would be considered low load. There is also nearly always 1GB of free RAM and disk IO is kept at a minimum. Looking purely at the resource usage, the system is not even close to being overloaded.

As a note, you may have experienced issues over this past week. We had announced a bug with Litespeed last week and that issue has since been resolved. Things are currently operating smoothly and should continue to run smoothly.

Thanks for the information Karl, How do I check the resource usage and the average load via SSH. Please let me know.

Regards

[Justec]

Thanks for the information Karl, How do I check the resource usage and the average load via SSH. Please let me know.

Regards

uptime




Yes, this is a short message Vbulletin

[Karl]

Thanks for the information Karl, How do I check the resource usage and the average load via SSH. Please let me know.

Regards

For resource usage levels, etc. you can just run "top"

[The Universes]

Web5 has been pretty hard hit this month and monitoring indicates its down right now as well. I don't use it to host anything mission critical or resource intensive, and thats not the purpose of shared hosting for that matter, so I'm not too concerned.

My only criticism is how Steadfast has handled this situation, it definitely could have been bettered handled and it just doesn't seem like the type of thing steadfast would do. I was a VPS customer here a few years ago and now I have a client with a dedicated server here. I've always remember support here being one of the best in the industry.

(Apparently I can't post links, but here is a screenshot of pingdom 1 minute monitoring alerts I've gotten this month: [add httpd] http://img260.imageshack.us/img260/3117/monitoringym0.png)

**Edit to make link work...**

[Ray]

I'm not sure of the exact situation with web5, but all of my domains are on that server, and I haven't noticed any issues since the crashing issue was fixed by LightSpeed. That said, I haven't been monitoring it or accessing it constantly enough to be able to say that it hasn't had any issues.

What sort of issues were there with handling the issues? I noticed a few announcements regarding problems, which I would think would be expected. I didn't contact support about the issues, because I was informed of the situation. I don't mean to suggest that what you're saying isn't true. I'd just like to know how you feel the situation could have been handled better, both for personal reasons and so the Steadfast support staff can improve in case of similar future problems.

[The Universes]

I'm not sure of the exact situation with web5, but all of my domains are on that server, and I haven't noticed any issues since the crashing issue was fixed by LightSpeed. That said, I haven't been monitoring it or accessing it constantly enough to be able to say that it hasn't had any issues.

The issue that occurred earlier today only affected my account because my site is on a "dedicated IP" and somehow the IP became "unbound" from the network interface. Unfortunately, it didn't get fixed until I contacted support about an hour and a half later, although I definitely should have opened the ticket earlier.

The only reason I have a pingdom account is because I use it to monitor servers for clients and I have a few slots left so I added it.

[Karl]

The issue that occurred earlier today only affected my account because my site is on a "dedicated IP" and somehow the IP became "unbound" from the network interface. Unfortunately, it didn't get fixed until I contacted support about an hour and a half later, although I definitely should have opened the ticket earlier.

The only reason I have a pingdom account is because I use it to monitor servers for clients and I have a few slots left so I added it.

OK, I had initially been confused, as the issues with web5 had been sorted out over a week ago. This is just evidence that if you're having problems, please contact us. We cannot possibly be able to help you if we are not made aware of the issue. Just as with this case, where it seems the issue was repaired shortly after we were contacted.

This is not a one way street, we do not know what the experience of each and every customer is. That means contact us if you have problems, contact us if things are going well. The more information we have, the better we can serve you.

[phaselight]

Same problem. Just contacted support. Mine has been down for 7 hrs and 20 mins and it's on web5. It's on a dedicated IP as well.

[Ray]

That means contact us if you have problems, contact us if things are going well. The more information we have, the better we can serve you.

Indeed. It's a lot easier to fix problems knowing that they exist. As much as I'd like there to be a way to constantly watch every single account with monitoring, it's just not feasible.

[The Universes]

OK, I had initially been confused, as the issues with web5 had been sorted out over a week ago. This is just evidence that if you're having problems, please contact us. We cannot possibly be able to help you if we are not made aware of the issue. Just as with this case, where it seems the issue was repaired shortly after we were contacted.

This is not a one way street, we do not know what the experience of each and every customer is. That means contact us if you have problems, contact us if things are going well. The more information we have, the better we can serve you.

Yeah, I understand, the issue wouldn't be obvious on your end and I don't think it was SF's fault.

However, the migration issues on web5 (which happened 1-2? weeks ago), which I was referring to, is.

Edit:
Maybe SF can setup some sort of hourly checks that make sure the dedicated IPs are bound properly and working.

[Karl]

As was posted in the announcement, https://support.steadfast.net/index.php?_m=news&_a=viewnews&newsid=175, the issue was actually a bug in Litespeed which happened to crop up at the same time as the migrations. We had initially assumed the issue was with the migrations, but the problem seemed to continue even when migrations weren't being performed. The issue was that when Litespeed was trying to gracefully restart it would just crash, etc. Without that issue, there should have been no real noticeable downtime from the migrations.