Dacentec - Unreliable Network Ran by Amateurs - Recommendations

[DomainBop]

Not to derail my own thread But....


Delimeter face punch today has me chuckling...  That's some karma for MarkTurner right there....


Refugee offers are nasty and I've created a number over the years for companies.  Refugee deals on physical servers, yikes, with that audience, that's going to hurt. Welcome to the microscope.

The most interesting thing about their 12+ hour outage is what the forum post counts about the outage reveal about their marketing focus and customer base: 348 posts and rising on LET, 0 posts on WHT.  


Goes back to waiting for that promised new microloan funded datacenter in Stockholm

 

[mikeyur]

348 posts and rising on LET, 0 posts on WHT.  

LET just seems to scream the loudest

 

[jarland]

So drmike, I signed up with Psychz today because I've been watching their success lately among VPS providers on LE*. Have you looked into them any? I don't know about you but I like a little DDOS protection in my network, and they're in the Carrier-1 facility in Dallas which is a pretty nice facility... same one Incero just moved to.


Maybe worth checking out? They've got some decent deals.


https://www.psychz.net/special-psychz.html (I had to open a ticket to actually get the listed deals)

 

[drmike]

So drmike, I signed up with Psychz today because I've been watching their success lately among VPS providers on LE*. Have you looked into them any? I don't know about you but I like a little DDOS protection in my network, and they're in the Carrier-1 facility in Dallas which is a pretty nice facility... same one Incero just moved to.

Yeah Psychz is on my short list... last go round pricing wasn't where I wanted it / shopping at.  I am still on fence with them and running a clean network, as they have taken in a lot of unsavory providers and had a fundraising method being friendly to mailers in the past.  Wait and watch with them, with no disrespect meant - just little old me has given free ride tickets in the past prematurely and taken it on the chin.


Now for Delimeter.... I am failing to understand WTF happened in ATL.  No one else in 55 Marietta had an issue with power that I can find.   A bird questioned if Delimeter has UPS units and gensets in place. Smells like they don't.  Not like people are paying to have such, just that Delim I think prior said they had all that stuff.


My experience in DCs is a bit, but not a DC engineer by any means.   This HVAC stuff + power is a weird ass combo.   Both break at the same time when utility power goes down.  Either a brownout with lame gear inline on the power conditioning or none.  Or worse, the utility drops the line for whatever reason, be it invoice overdue or some slob with a jackhammer. No slob seen.


And back to HVAC, the overheating and conditioning clearly indicates HVAC tonnage isn't aligned with the BTUs put out by that dense blade gear / all that gear in the space.  I've seen this multiple times in the past.


And the power tiering of bringing units up - Delimiter is over loading circuits period.  No reason to tier like this to get gear back online.  Other than trying to prevent fuses from popping and the electrician from having to deal with humming deads man switch when the whole buildout pops for over current.


I am inclined to run a fine toothed comb over this and contact everything and everyone in 1 block area from 55 Marietta.  Cause my spidey senses are ahh tingling.


Stuff doesn't add up on this one.   Call me names and tell me I am full of it, been in DC environments since 1990's.  Seen too damn many stunts.

 

[GM2015]

Servers are down for more than 1 day now. I hope I won't be sorry for renewing my server with them.

 

[jarland]

Wait and watch with them, with no disrespect meant - just little old me has given free ride tickets in the past prematurely and taken it on the chin.

I feel ya.

Now for Delimeter.... I am failing to understand WTF happened in ATL.

To me it sounds like a major failure of their colocated equipment combined with remote hands. We all know well how inefficient remote hands can be at times, and often having no sense of urgency or care for the impact of what they're doing. Anyone taking mild offense to that is quite likely an exception to that so they shouldn't


Edit: I stand corrected, according to Mike they have direct employees there.

 

[drmike]

I feel ya.


To me it sounds like a major failure of their colocated equipment combined with remote hands. We all know well how inefficient remote hands can be at times, and often having no sense of urgency or care for the impact of what they're doing. Anyone taking mild offense to that is quite likely an exception to that so they shouldn't


Edit: I stand corrected, according to Mike they have direct employees there.

They claim to have staff there on location.. they claim to have own build out.. claim to have had power failure on power transfer switching was it and the HVAC....  


Shit happens, but that was an unreasonable amount. Karma bitch slapped them either way.


I fully expect to see a DC tour and busted stuff... after all that drama and lost revenue and shredding of reputation.   Lots of us don't believe the build is what it is.   Posting some real stuff would put many at ease. @mikeyur see what you can do.

 

[DomainBop]

I feel ya.


To me it sounds like a major failure of their colocated equipment combined with remote hands. We all know well how inefficient remote hands can be at times, and often having no sense of urgency or care for the impact of what they're doing. Anyone taking mild offense to that is quite likely an exception to that so they shouldn't


Edit: I stand corrected, according to Mike they have direct employees there.

30+ hours downtime = major failure by Yomura/LTT management in planning/design/running of their data center.  The equipment failure they had should not have resulted in a 30+ hour downtime for many customers. The hired "remote hands" at Delimiter (or on site techs) weren't responsible for cutting corners on cost and building a DC with no power redundancy, crappy DC infrastructure equipment (just like, getting back to the original topic, the remote hands at Dacentec weren't responsible for that unreliable network with no redundancy or for CentriLogic's  monumental flop Mothership1).  


The decision by hosting companies to write a business plan that includes relying on crappy inefficient 3rd party unmotivated workers to cut costs (which is often a necessity because the companies had the bright idea of engaging in a race to the bottom on pricing), whether they're remote hands in a DC or outsourced poorly trained overseas labor (or in some cases 10 yr old migration specialists) also rests entirely with the management of the companies.

 

[drmike]

@DomainBop

Ever realize that Mothership1 was another crafted and outsourced creation of a Europe boutique specializing in creating these 'cloud' companies:
http://incubaid.com/company/About%20The%20Company/

Dacentec: http://incubaid.com/successes/Dacentec/

Mothership1: http://incubaid.com/members/Mothership1/

 

[mitgib]

rests entirely with the management of the companies.

Does it? Or does the fault rest with the clients? Companies are budgeting what the client is willing to pay.  Not everyone is willing to pay $1k/mo per cabinet in a properly built facility, and when the customer cuts corners on cost and does suffer the eventual failure, why is the DC to blame for giving the customer what they wanted?


Also, as the story goes, shit happens

 

[drmike]

I am of the unreasonable school of belief that goes like this... I expect to get crap for this [rant and ramble mode ON]:

* Customers don't set the prices, the seller does in order to sell and attract customers.

* In the cheap world, companies are notorious for trouser drops to make sales. Most have no other magic to sell on. It's self robbery / disgrace or sales volume won't happen.

* Customers expect what is advertised / marketing puts out there

* In Delim's example they boasted in REFUGEE ad about all this power redundancy:
"... Every server has multiple power feeds, blade chassis are fed by two independent three phase supplies across 6 separate power modules"
^ This is either fluff or they forgot more important redundancy. Seems like they had everything single legged on power.

* More fluff / lie about quality of gear / age of drives
"... Every server comes with new or as-new disks, we do not use 5 year old second-hand or refurbished disks ..."
^^^ at least one customer posted data from his drive with over a year of run time on it. Second hand and not new. Violates at least two of the sales BS points.

* Every server is a branded HP server, not some hand-me-down from the latest Ebay liquidation
These are 54xx and earlier CPUs. No one is selling this new or willing to buy or be stupid enough to buy legit new units at full retail price for those at this point. Therefore these are off lease or if Delim bought them new, they did so many years ago and have used and beaten the life out of them in this time. White lie maybe, total lie maybe.

* Customers should never be lied to or falsely sold. This all happened with Delimeter. Hell doing so is technically against multiple laws / regulations.

* Even if customers received services for free, there is expectation level that barely is diminished when it comes to sitting in unknown with data offline. Free isn't even a valid legal argument for shit service. Service is power + ping + support delivered at level that is advertised or in lieu of, common in competing shops.

* Price doesn't equate to service delivered. There are plenty of expensive shit co's overloading things and failing horribly just the same while making a lot more money.

* Price point only equates with expectations where a company says HEY THIS IS A WARNING. WE HAVE LIMITED STAFF. WE USE OLD GEAR. THINGS ARE OVERSOLD.
^^^ remember Delim sold people on pipe smoke dope and hype of being a super bad ass company with superior network, great uptime, all this high end up... That's the fatal mistake.

* Never provoke Karma's Wrath or attract Murphy's Law.
^^^ poking competitors hard with a sharp stick is just bad. If Dacentec failed ugly and booted customers, sure refugee it. Dacentec got DDoS'd 190x probably with blips (why I am running away from them). Sure they have flawed power setup too. But downtime was meh. Only irritated intolerants like me cared enough to bail. I've been on about bailing for months prior to uptick of issues in Lenoir.

* CUSTOMERS DON'T DECIDE PRICE POINT, THESE COMPANIES DO
--- I have yet to see but a handful of companies honest enough to say they are understaffed, delays in support, old gear, yada yada... Every monkey in this industry screams about enterprise and what super bad ass 15 second support they have. As if they employ 10+ ADHD support agents with an electro shock jock strap that forces them to spring into action when ticket comes in. Utter tripe all of it, 70% of the companies are full of shit and always have been.

Is Delim's support bad? Sure, I think it is utterly random at best historically. Shop LET threads and see people waiting days for stuff and not just dedi delivery. Common support matters.

I am so tired of the shit support all over that I am contemplating starting a support company. I could train literal monkeys to do better than 50% of the companies out there do on support.

[/rant mode off]


PS:  I felt @mikeyur handled LET well and kindly in light of distance and probably working from limited knowledge / familiarity of that location / limited resources.

 

[mitgib]

I expect to get crap for this

 

[sv01]

how about admin/mod move delimiter downtime to it's own thread?


WHT Link : http://www.webhostingtalk.com/showthread.php?t=1547582&p=9625565#post9625565