Media Temple Customers

A blog by (mt) for (mt) customers

Posts Tagged ‘Downtime

System Status Ambiguity

with 10 comments

I keep getting comments on this site and Twitter messages about sites being completely down or in various states of disarray. Some customers complain that Media Temple has not called them on the phone or explained anything.  Sites being down this long is insane and the status page does not really clarify things:

recoverystatus

Media Temple should definitely communicate better with these users. All that is gained from the Status Incident Page is that sites are being moved at a very slow pace. Media Temple should answer these questions to the customers who are still experiencing problems:

What does this mean? 

Why were some sites back online yesterday and others still offline right now?

Why is it taking so long?

What is this thing about the Fiji Agent? Can I perform migration on my own to speed things up?

Was the multiple phone calls I got due to my loud complaints on this blog and Twitter? I don’t know. For me 8 hours of downtime, 20 hrs of Account Center lockout and intermittent email/ftp errors which translated into 1yr of free hosting seemed fair… It’s like beta testing their new architecture… as a web developer free hosting to play with is something I would always agree to even if it’s unstable (which I hope the new Cluster will be)

My personal sites and projects I plan keep on this Media Temple (gs) account. Why would I trust them at this stage? Primarily because an account I have on the newer Gen 2  architecture on Cluster 4 hasn’t experienced any serious downtime and I’m led to believe my migrated Cluster 2 sites will be fine from now on out.

My clients who squawked the loudest I will be moving to a different ISP. I have a responsibility to them and if they would feel safer that way, I will oblige.

For people whose sites are still down and losing Alexa rankings, traffic, ad revenue, and clients- that really sucks. I don’t know what can be done for them. This is my attempt to help, as a small web developer  I understand your frustration. Please leave suggestions here, perhaps your voices will be more unified than on Twitter. Media Temple does seem read all the comments and posts on this blog very quickly.

Advertisements

Written by bnvk

May 6, 2009 at 10:05 pm

Cluster 2 Back, Customer Care, Migration Recap

with 6 comments

So Cluster 2 was back online after 8+ hours of downtime in my case, and still down in some other  peoples cases. My Account Center is still locked out after 19+ hours and some sites of mine have errors that I can’t seem to troubleshoot due to the Account Center Lockout. 

picture-2

I got a phone call from Andrew, a Customer Service V.P. at Media Temple whom I have spoken to before, here are some points that were addressed.

“What percentage of Cluster 2 customers have been migrated to Gen 2 architecture?”

Andrew: “44% of Cluster 2 Customers have been moved.”  (which is the same as it was back on March 6th)

Brennan: “Wow!”

Andrew: “We just finished our Fiji Agent last friday. This will speed up the process.”

“Why was no one moved during the last two months? Or at least why was whiny little bitch customers like me not moved, even after I asked to be moved after the 36 hours of downtime?”

Andrew: “We had faith that the Cluster 2 hardware would be ok.” 

Brennan: “Seriously? You had faith that after 36 hours of downtime [the same hardware setup] would be ok?” 

Andrew: “Well, let me rephrase that. We wanted to make sure we were making the best decisions to migrate people at the best times so there would be minimal amounts of downtime” …and… “we have been putting all of our resources and developers into  the Fiji Agent.”

Brennan: “Wow! Too bad this had to happen.”

Media Temple claimed to be on a “road to better transparency” something about this doesn’t feel transparent to me. I said “Transparency doesn’t mean talking more, it means being honnest”

Andrew: “We haven’t been dishonest about anything.”

While technically that might be true. Media Temple did claim on March 6th that:

Next 30 days, Cluster-to-Cluster migration tools. We have committed significant developer and administrator manpower to the development of Cluster-to-Cluster migration tools.”

I understand. Software development does sometimes take longer than anticipated…

But for some reason I am inclined to say bullshit. If it looks like a bull and smells like shit it’s probably bullshit. I seem to remember around the time of SXSW (March 17th), during the absurdly long period (1.5 weeks) of POP email server IP blacklisting, there was a lot of Tweets about how amazing the Media Temple SXSW party was. In fact they threw The Official Closing Party for the SXSW Interactive Festival. This party was replete with international music groups, an open bar, a photo booth, and a mechanical bull. 

sxsw_bull

Maybe I’m off base here, but I imagine throwing a party of that scale takes time to organize and a substantial budget to pay for. Oddly, I don’t see the correlation between mechanical bulls and fixing unstable web servers but maybe that’s just me.

In hindsight, considering this new downtime debacle, and the facts, I have to ask the question: How much of Media Temple’s resources were devoted to getting frustrated customers to a more stable hosting platform like their Gen 2 architecture? 

Oh, Andrew also offered to move my sites within the hour to Cluster 4 so I could use the Account Center… this however could result in more patches of downtime… I wonder if they understand that downtime during the middle of the day is a bad thing for websites? I said “How about you do that later tonight.”

**The quotes may not be exact as I didn’t have an audio recorder handy. I tried my best. 

Written by bnvk

May 5, 2009 at 6:23 pm

8+ Hours Of Downtime

with 13 comments

After 4+ hours of unexpected server downtime, a couple irate clients, 1 semi angry Tweet I received a call from  customer service at Media Temple. She wanted to know if I had any questions…

My only question was:

“What percentage of Cluster 2 customers have been moved from Gen 1 architecture to Gen 2 architecture”

This customer support lady didn’t know the answer to that question. She put me on hold to try to get the answer. After a few minutes waiting she said “they have no way of knowing exactly.”

That was the wrong answer. Because now I am left feeling that Media Temple either has no clue about what they’re doing or they are not being honest. How can I jump to this conclusion? Well on March 6th 2009, following the 36+ hours of downtime, Media Temple made this blog post: Anatomy of Storage on the GRID

It explained all the old architecture (Gen1 ) and new architecture (Gen 2). 

2gens

This blog post also claimed:

“At time of writing, 37% of Cluster 1 customers, and 44% of Cluster 2 customers have already been migrated.”

So I am forced to really wonder how 2 months later they don’t know at all… I mean 80% or 70% percent would have been undesirable answers but “We have no way of knowing” is kind of absurd. They didn’t project to be done with until June but what I am left to think now?  

In either case I  getting very very distrustful with them as a hosting company. I was one of the customers affected by the 36 hours of downtime staring Feb 28, 2009. I had setup this blog. I got a lot of attention for rallying the twitter/blog-o-sphere. Back then I got a phone call from a V.P. at Media Temple followed by personal phone number for V.P. at Blue Arc (mt hardware supplier). They both sorta blamed the other for the downtime. Fair enough. It still sucked. I didn’t want to go through the headache of moving to a new ISP and I was appeased by some PR people- they gave us all free hosting for a month. But now another HUGE outage is just unacceptable. 

 

 

Written by bnvk

May 5, 2009 at 4:09 am

Collection of Media Temple Customer Experiences

leave a comment »

Here are few links to well written informative posts about how others are experiencing this horrid amount of downtime. 

Duane Storey – The Day That Media Temple Stood Still

Steve Reynolds Blog – An Open Letter to Media Temple and Media Temple Outage affects Customers

Social Marketing Contract – Media Temple Provides Religious Sunday Experience

Karelia Software – Weekend Website Outage

And there is always the live Twitter Search

Written by bnvk

March 2, 2009 at 9:34 am