A lot of people are upset about the large scale outage that caused many sites hosted at EIG controlled companies like Hostgator and Blue Host to go down yesterday. One post I read today suggested that a move to LiquidWeb is the solution.
Having hosted with Hostgator for 4 years I have found their service cost effective and reliable with adequate support. Fortunately, both our Hostgator servers are located in their Texas data center so we were unaffected by the recent network problems at Provo.
I can empathise though with the sentiments expressed by the person advocating everyone move to LiquidWeb. There is nothing more frustrating than to have your sites go down and not be able to do anything about it. Companies like LiquidWeb will find they have a large number of new clients in the coming weeks.
However I am not going to be so quick to make the decision to dump Hostgator and move back to LiquidWeb, as we were in a similar position around 18 months ago at LiquidWeb!
Not So Smart VPS Servers
We had been happy LiquidWeb customers for 2 years using a standard or legacy VPS system. Their support was bordering on the heroic. Then in September 2011 they told us we’d need to move off our Legacy VPS and on to their new SmartVPS systems. We were not keen on moving our server as it involved moving a number of sites plus a web application. However, there was an issue with the lifetime of the Centos 4 operating system. LiquidWeb said :
On February 29, 2012 The CentOS Project will officially stop support for CentOS 4. After this date no further security updates or bug-fixes will be released.
Our only option was to upgrade to the SmartVPS.
We migrated to the SmartVPS in November 2011. And that’s when the hacking attempts and performance problems started. With the move we got a new set of IP addresses, and it was as if those IP addresses were jinxed.
LiquidWeb SmartVPS Servers Left Us Smarting
After the move we had numerous ongoing issues. The system was being attacked repeatedly and performance was in the toilet. LiquidWeb seemed incapable of blocking the attacks and even suggested it was our fault. A real LiquidWeb human may pick up the phone after a couple of rings, but there were times when a recorded message would have been preferable to the answers I got. Their solution? We should upgrade to a more powerful server, taking us from $90 per month to $150 per month.
The attacks never stopped – we (no help from LiquidWeb) would fight back and the attacks would subside for a couple of days. During this time we were only using 5-20% of system capacity so upgrading did not make sense to me. What we needed was a way to block the attacks to prevent them consuming 100% of CPU and memory, leading each time to total systems failure.
Superhero Support? No, More Like The Joker
No-one – and I mean no-one at LiquidWeb had neither an explanation nor an answer. No-one helped us despite my pointing out the obvious – that the problem had started the day we moved to the new server. Before that, no attacks. The day we moved the attacks started. When we asked for new IP addresses they REFUSED. I had to research and address the problems on the server that I had not created. Where was the heroic support when I really needed it?
We were getting two or three major outages each month that lasted hours. It was like clockwork. Our clients were becoming alarmed with the frequent downtime and poor performance and we spent several sleepless nights each time – making sure the servers were back up in time for when our UK customers woke up! In addition to this LiquidWeb had a couple of complete network outages in January and March 2012; even their ticketing and support systems were offline for several hours.
Eventually, in April 2012, after many calls and hours of frustration the problems persisted and in order to stay in business, I had no option but to move all my sites away from LiquidWeb and on to a stable and secure platform on Servint.
Servint have been excellent so far. It’s like being with LiquidWeb back in the gold old days. He he.
I note that now LiquidWeb have replaced their SmartVPS platform in favour of the StormVPS platform. I am more than willing to take a look at them again as an alternative to Hostgator.
However it is not entirely clear to me what the difference is between SmartVPS and StormVPS other than the pay-as-you-go bandwidth costs? It is just a re-branding exercise or have they solved the security issues with SmartVPS by introducing better management software? I hope so.
So What’s The Moral Of This Story
In this game nothing stands still. And there is no magic hosting solution.
The need is to be flexible. Stick with your host in the good times and if the pain becomes too intense and they can’t turn things around then by all means, move on. But please know, the only true answer lies in resilience. Swapping Hostgator for LiquidWeb means you may have just swapped your single point of failure for another.
The answer for us has been in resilience. Where we had one server before, now we have five, so the impact of any outage is much reduced.
I have also written failover software to ensure that my web applications stay up which is what matters to me. When one server fails, the next one (hosted somewhere completely different) comes in automatically. I have layers of 5 failover servers which do that job and I can add as many I need. Now I can feel a bit more comfortable.
Also by having one of our servers in the UK we have time to do out of hours maintenance should the need arise without affecting US clients. Resilience has its benefits.
Checking Out LiquidWeb StormVPS
While Hostgator may be on a downward path as EIG try and squeeze more out of it, LiquidWeb may be on the up. (Let’s face it, the only way was up from our point of view). I will be renting a StormVPS in the coming week.
I want to check it out in depth so I know for sure that WP Whoosh can deploy to LiquidWeb StormVPS and hence I can add it to the list of recommended WP Whoosh hosts.