Webloggin » Update: Another Sign That 1and1 is Simply the Worst Web Host Imaginable

The 1and1 Web Host Debian Etch Debacle Continues

Update : Turns out I called support and they were unavailable when my server was down again. The conversation went something like this:

Girl from some foreign place, assuming India 1:00 a.m cst - : I’m sorry sir but the Server 24×7 support team is on a break and they are the only people who can reboot your server at the moment.

Me: !! They are on break? Well when will they be back so I can get my server back on line? How long is the break?

Girl from some foreign place, assuming India: 3 hours sir.

Me: Excuse me, I think you said 3 hours.

Girl from some foreign place, assuming India: Yes sir, 3 hours.

Me: So you mean to tell me that my dedicated Server 24×7 support team, the only people that can reboot my managed server, is on break for 3 frickin’ hours??

Girl from some foreign place, assuming India: Yes sir.

Me:

Girl from some foreign place, assuming India: Is there anything else I can do for you?

Me: [Not very nice statement of facts occurred here.]

Girl from some foreign place, assuming India: Thank you for calling 1&1.

Original Article Below:

My 1and1 Hosting Nightmare or “Why Should I Keep My Business With a Server/Web Host that Suddenly Sucks?”

It is unfortunate that the sad state of performance and technical support from my server host has compelled me to write this article. Consider this a warning to others that are looking to pay premium dollars to have a dedicated server hosted on a managed platform. If 1and1 is in the running and you need any sort of guaranteed uptime with guaranteed support then look elsewhere. BTW, if you are reading this and you are a server host or have a reliable server host please email me

For the record I have been with 1and1 for a few years without issue despite the warnings of others who have obviously had the same terrible experience that I am dealing with today. On average we have incurred about 3 outages a year. That is until recently where we have been hit by a spate of outages; incurring more downtime in the last month than in the whole year before. This affects Webloggin and many other domains that I host; people with businesses that just need a reliable host.

This is where 1and1 becomes simply atrocious.

The first shortcoming of 1and1 is their level of support. My contract comes with what is considered tier 2 support on a 24×7 basis:

Dedicated Server Support

If you need help with your server, please contact the 1&1 Dedicated Server support team.

Phone 1-877-206-4255
E-mail server247@1and1.com

Now I don’t know about you but in the english speaking world this sort of implies that there will be a support tech available to help you when your server does something terrible; like spinning CPU at 100%, refusing to serve up pages, shells or any of the normal things that I pay to have available to me and my customers.

So imagine my surprise when my server was down again for the third time this week and I was routed to technical support. Of course I didn’t know this at first because it took a while to drag it out of the support tech but I have learned that just about anytime I get routed to a foreign location, say perhaps in India, that I am not at the server support line.

My conversation with the Indian support crew was less than satisfying for either of us. Being irate at having the server off line yet once again I demanded 1.) an immediate reboot and 2.) an explanation of what the hell was going on. The tech responded that he would have to contact the server support team to request a reboot. But wait, I called the server support line! The tech said that calls get routed to them whenever the server line doesn’t pick up. Please hold sir.

After 5 minutes of listening to the same crappy 80’s rock that I have experienced over the last couple of years the tech came back. I’m sorry sir, but the server support team is not answering. We will send them an e-mail with our highest level support team so they can address this as soon as possible, he says.

At this point I essentially lost it. You mean that my dedicated support is essentially a stenographer that will send an e-mail to the other server support team that may or may not address the issue in the next 24 hours??? Sorry, that is completely unacceptable.

I’d like to speak to your manager.

Please hold sir while I check to see if my manager is available to take your call.

Hold with same crappy 80’s rock.

I’m sorry sir, but my manager is not available to take your call.

You get the point. I get a case number and no resolution. So I give it a few, probably 5 minutes and call back again. This time I get an Indian girl. I explain the situation and ask for more help. I am sorry sir but the server 24 support is closed. We will send them an e-mail to see if we can get your server back up as soon as possible.

What? You know how to claim dedicated support but you don’t know how to give it!

On the third attempt I finally reached the server support line. The tech informed me that the issue was with a recent operating system upgrade to Debian Etch and that a system process that runs on etch has resulted in 95 consumption of all cpu. This of course is after a tech told me 2 days ago that Wordpress was causing the database to hang on long running queries. (Thanks, I optimized all my databases after that call). I also upgraded all my PHP based domains to a more recent version. I ran a top an essentially my user processes are nearly idle with plenty of memory available. Root processes are however killing the box.

To make a long story short. My server experience is suffering because:

the recent upgrade to Debian etch has been detrimental and nobody bothered to fess up until today.
Server 24×7 support doesn’t mean crap. It’s an untruthful selling point IMHO.
It feels like 1and1 has outsourced technical support. This essentially means that I may as well be managing the server myself; kind of counter to the point of having a managed server in the first place.
1and1 doesn’t care to keep my service. If I leave there will be another sucker down the line to back fill my business.
Today’s explanation doesn’t come close to making sense. As you can see at the bottom of the page my CPU snapshot does not show any spike in CPU as has been characteristic in the past. I believe the upgrade to Debian has also affected the monitors, just another broken thing on my premier server.
1and1 doesn’t respond to my e-mails, doesn’t have a customer support number that I can find, and didn’t even set up the debug logs I requested that were promised to me.
I have repeatedly asked 1and1 to tell me what is causing the server outages and have never received a satisfactory answer.

Thanks 1and1, for nothing.

Typical top capture:

top - 20:26:53 up 1:19, 2 users, load average: 0.45, 0.64, 0.55
Tasks: 27 total, 1 running, 26 sleeping, 0 stopped, 0 zombie
top - 20:37:49 up 1:30, 2 users, load average: 0.55, 0.51, 0.50
Tasks: 23 total, 1 running, 22 sleeping, 0 stopped, 0 zombie
Cpu(s): 60.1%us, 10.6%sy, 0.0%ni, 28.6%id, 0.0%wa, 0.3%hi, 0.3%si, 0.0%st

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
7532 XXXXXXX 15 0 53164 6044 1752 S 1.7 1.2 0:01.14 apache-perf
3820 XXXXXXX 15 0 53168 6028 1736 S 1.0 1.2 0:01.39 apache-perf
7803 XXXXXXX 15 0 53172 6036 1732 S 0.7 1.2 0:01.15 apache-perf
……

(uiserver):XXXXXXX9:~ > vmstat
procs ———–memory———- —swap– —–io—- -system– —-cpu—-
r b swpd free buff cache si so bi bo in cs us sy id wa
0 0 0 92824 33504 269708 0 0 54 71 285 182 16 4 79 2

As you can see my user processes are essentially idle. Hence the root processes are consuming most of the CPU and it comes down to some sort of monitor.