GoFuckYourself.com - Adult Webmaster Forum - View Single Post

thickman · 09-30-2002, 02:20 AM

Cologroup,

Oddly enough I find myself posting to this thread. My name is Ted Hickman, as most know I am the President of another hosting company, I post as a fellow hoster and not as a competitor.

This post actually touched me, first because of the
harshness that consumers show towards their hosts. We too had virtual server problems before clusters and after. Our virtual cluster was down for 3 days BECAUSE we did a cluster, and later found to be bad drive connectors in our RAID units. Our customers stayed, and were credited appropiately and that is the most they know. What they don't know is that I and two of my guys slept on the floor of the datacenter for 3 days while this was going on. I sacrafice for my customers as I am quite sure you do by your posts. However that is the business we are in, we made the choices we need to learn to deal with the lack of appreciation. After the last downtime, I learned I would rather raise my prices and charge and maintain for a premium service then charge absolutely ridiculous prices with which I could never maintain my service. I would much rather be a National Net type of operation than a dollarpergig.com type of hosting.
And I am striving towards that goal everyday.

About clustering, let me first say it's very difficult to do and expect lots of problems. It took me nearly a year and half to iron out clustering for our virtual clients. F5 had a problem with someone who had 40 thousand ips used as vips on their BIGIP's, the user management fell out of sync. The ability to provide logs and a good stats program plagued us for a LONG time, NFS And redundancy also became a huge issue with replication costing 2k per server with 15 server cluster and 4k for the raids themselves. I and my team came up with solutions, as fast as possible, and now have a very good operation in place. Trust me I have been through the clustering nightmares and there are a lot. Just don't kill yourself if you can't get it right the first time, you will get it right, and customers will come and go as they have with our company.

On the subject of redundancy and failproof systems, they simply don't exist, redundancy CAUSES a lot of issues. Infact, 99% of the issues with the CLUSTER has been with the redundancy and would not have occurred if it were not redundant. Redundancy is a very sensitive fragile thing, any provider that uses two of anything exactly the same is in for trouble. DOS, natural bugs etc will hit the same devices the same ways every time.

I hope that in some way I have helped by this post to let you know that other hosting companies go through this sort of thing. WE all do, whether we admit it or not. OUR industry as a whole goes through these problems. Qwest isn't even selling virtuals anymore because of the headaches. Likewhoa getting shut off by VERIO is another classic example of problems hosts have and how they deal with them. This I think in the end is what customers base our business on, service and how we deal with the inevitable problems that will happen. And yes there are some customers that will never be reasonable no matter what you give them or how you serve them.

I know how frustrating it can be to build something up just to have someone or something knock it down, but it does and will continue to happen in the hosting industry.

Lastly, I think it's a cut-throat business, everyone jumping down everyone else's throats, vultures picking at your body before you are even dead, however what goes around comes around and I am quite sure their problems are lurking around the corners, and they will soon learn what pain it is that you feel. The thing that bothers me the worst, is the dozen's of tutorials or posts out there by people that have never run a hosting company trying to illustrate and instruct on how to run one. They think it's as simple as bandwidth and servers, I can tell you from experience it's not. I applaud you for being honest and direct in your posts and making progress in fixing the issues that plague you where most are content to laugh or make fun of others, before looking at their own.

I have said my piece, please no flames.

Ted

09-30-2002, 02:20 AM
thickman Registered User Join Date: Sep 2002 Posts: 83	Cologroup, Oddly enough I find myself posting to this thread. My name is Ted Hickman, as most know I am the President of another hosting company, I post as a fellow hoster and not as a competitor. This post actually touched me, first because of the harshness that consumers show towards their hosts. We too had virtual server problems before clusters and after. Our virtual cluster was down for 3 days BECAUSE we did a cluster, and later found to be bad drive connectors in our RAID units. Our customers stayed, and were credited appropiately and that is the most they know. What they don't know is that I and two of my guys slept on the floor of the datacenter for 3 days while this was going on. I sacrafice for my customers as I am quite sure you do by your posts. However that is the business we are in, we made the choices we need to learn to deal with the lack of appreciation. After the last downtime, I learned I would rather raise my prices and charge and maintain for a premium service then charge absolutely ridiculous prices with which I could never maintain my service. I would much rather be a National Net type of operation than a dollarpergig.com type of hosting. And I am striving towards that goal everyday. About clustering, let me first say it's very difficult to do and expect lots of problems. It took me nearly a year and half to iron out clustering for our virtual clients. F5 had a problem with someone who had 40 thousand ips used as vips on their BIGIP's, the user management fell out of sync. The ability to provide logs and a good stats program plagued us for a LONG time, NFS And redundancy also became a huge issue with replication costing 2k per server with 15 server cluster and 4k for the raids themselves. I and my team came up with solutions, as fast as possible, and now have a very good operation in place. Trust me I have been through the clustering nightmares and there are a lot. Just don't kill yourself if you can't get it right the first time, you will get it right, and customers will come and go as they have with our company. On the subject of redundancy and failproof systems, they simply don't exist, redundancy CAUSES a lot of issues. Infact, 99% of the issues with the CLUSTER has been with the redundancy and would not have occurred if it were not redundant. Redundancy is a very sensitive fragile thing, any provider that uses two of anything exactly the same is in for trouble. DOS, natural bugs etc will hit the same devices the same ways every time. I hope that in some way I have helped by this post to let you know that other hosting companies go through this sort of thing. WE all do, whether we admit it or not. OUR industry as a whole goes through these problems. Qwest isn't even selling virtuals anymore because of the headaches. Likewhoa getting shut off by VERIO is another classic example of problems hosts have and how they deal with them. This I think in the end is what customers base our business on, service and how we deal with the inevitable problems that will happen. And yes there are some customers that will never be reasonable no matter what you give them or how you serve them. I know how frustrating it can be to build something up just to have someone or something knock it down, but it does and will continue to happen in the hosting industry. Lastly, I think it's a cut-throat business, everyone jumping down everyone else's throats, vultures picking at your body before you are even dead, however what goes around comes around and I am quite sure their problems are lurking around the corners, and they will soon learn what pain it is that you feel. The thing that bothers me the worst, is the dozen's of tutorials or posts out there by people that have never run a hosting company trying to illustrate and instruct on how to run one. They think it's as simple as bandwidth and servers, I can tell you from experience it's not. I applaud you for being honest and direct in your posts and making progress in fixing the issues that plague you where most are content to laugh or make fun of others, before looking at their own. I have said my piece, please no flames. Ted __________________ Ted Hickman