Hi, it always amazes me that if you ask a host when they last time they went down was, they always have a really bad memory.

Anyway, I just set up an ezzi server, and this recent downtime frightens me. I asked them how often it happens, and the support guy can only remember the New York blackout last summer.

So, can anyone give me an idea of how many times per year ezzi goes down? How long, etc?

Take a look at http://www.webhostingtalk.com/showthread.php?s=&threadid=324280&highlight=ezzi
you're being played

You are refering to the downtime yesterday of about 1 hour? No, I knew about that. I was asking when they had been down before that. They said it was over a year ago, and they couldn't remember any time earlier. They didn't try to lie to me about yesterday's downtime.

To be honest, stay away from ezzi if you want decent uptime for your servers. My last experience with ezzi last year was not a pleasant one. Their network has been down quite often compared to other hosts. The network is fast when it's up but it looks like it's not fully redundant. Take the last downtime as an example about the UPS crash. A decent and properly built datacenter should never have any such issue.

i'm currently on ezzi right now, i've been with them for about a year or two.. the network was bad back then but has improved greatly since. I wouldn't recommend it for mission critical data, however you can use it like i do.. for personal data. Ezzi is trying to improve their network and ive seen a big improvement.

Servermatrix had the same issue a few months back. It can happen. Hopefully not anymore. 4 servers down yesterday

EZZI's core network has been up 100% this year. We may have had local issues here or there, like what happen yesterday that was very unfortunate and also out of our control... but the core network has been up 100% this year.

Problems will happen, it comes down to how fast the provider deals with it and also how they communicate with the customers letting them know what is going on.

I have had servers at ezzi for a few months and yesterday was the only downtime that we have seen. The network is pretty good.

While no one likes downtime, yesterdays was handled pretty well and back up when they said it would be.

About a year ago, ezzi was known for having a terrible network. People always said stay away from ezzi back then. Then they kind of went low, kept a low profile while building and improving their network. And a few months ago they began a very strong comeback and have since proven themselves. They are about 1/2 the price of servermatrix, double the bandwidth, higher uptime, and support responses in UNDER an hour (not 6 hours+ like servermatrix). Although I would not consider ezzi to be extreme mission critical (like RackSpace mission critical) i do consider them to be absolutely superb in almost every aspect of their business.

Originally posted by RaineTech
About a year ago, ezzi was known for having a terrible network. People always said stay away from ezzi back then. Then they kind of went low, kept a low profile while building and improving their network. And a few months ago they began a very strong comeback and have since proven themselves. They are about 1/2 the price of servermatrix, double the bandwidth, higher uptime, and support responses in UNDER an hour (not 6 hours+ like servermatrix). Although I would not consider ezzi to be extreme mission critical (like RackSpace mission critical) i do consider them to be absolutely superb in almost every aspect of their business.


You are right, last year we grew way too fast and it hurt us a lot. We finished all upgrades about the end of last year and started pushing servers again I would say around the new year. Since then as everyone can see just by posts in this forum or even by what customers say, things have been great.

Uptime is great, live chat works well, we spent a ton on protecting the network from dDoS attacks and such.

There will always be those people that come on here and say EZZI sucks when they had servers with us over a year ago. Problem with that is, it isnt accurate, and credit should be given to a company that had problems, took a step back, re-vamped everything and actually kept their word on what they said they would do in terms of getting the network solid.

We are coming up to a year now and as I said before the core network has been up 100% of the time this year. We are very happy that things are going well and I am sure our resellers and customers are just as happy.

There are some companies that make promises that do not come through on them. We have come through on everything we promised to our customers last year and that means a lot to our long standing customers. Those are also the things that just dont make it to posts on any forum.

Even with all the money we spent on protecting the network, the complete over-haul on the network, bring in a new core network team, and all the other stuff that has been done around here... I believe it was somewhere around a million dollars... we are still known for being the most competitive in the industry.

We may not have the quality name of a Rackspace, but for the money you spend... we have many customers that believe the value of the service is much greater than many other companies in the budget market.

EZZI has also never taken any investment money to date. Since EZZI.net is part of the EZZI conglomerate of 7 companies and over 1,000 employees world-wide... it is a great feeling that our customers and resellers have when they know EZZI will be around. As we all know, in this industry, that can be a big problem.

Can I ask one question? What are your other companies are they in the hosting space?

Originally posted by ispclub.com
EZZI's core network has been up 100% this year. We may have had local issues here or there, like what happen yesterday that was very unfortunate and also out of our control... but the core network has been up 100% this year.

Problems will happen, it comes down to how fast the provider deals with it and also how they communicate with the customers letting them know what is going on. Is the core network measured at the core route at EzziCenter3000 (where all of your normal Ezzi servers are), or is it measured at your main NYC POP? There is a difference if the NYC POP had 100% uptime, but the uptime measured from the core router at EzziCenter3000 had 99.5% uptime.

Originally posted by Joshua
Is the core network measured at the core route at EzziCenter3000 (where all of your normal Ezzi servers are), or is it measured at your main NYC POP? There is a difference if the NYC POP had 100% uptime, but the uptime measured from the core router at EzziCenter3000 had 99.5% uptime.


The core network would be all main routers and switches in all of our locations together. Basically we have not had a network-wide outage this year.

In any local outage we have had, I would say the most amount of servers that have been down or had high latency for whatever reason was about 10%.

Originally posted by coolraul
Can I ask one question? What are your other companies are they in the hosting space?


The other companies do anything from consulting, hardware, VoIP, call centers, software development, and whatever else they have going on.

EZZI, how would you define your 'core network'? I saw your explanation above, but that would seem to be your entire network... not just the core. Anyway, just curious.

Do you separate the network for your high end customers and low end customers? If you said only 10% got affected by the UPS crash yesterday, are those all customers of your budget (low end) servers?

It's great if ezzi redundancy has improved much since last year and I may give it another try.

Originally posted by blueface
Do you separate the network for your high end customers and low end customers? If you said only 10% got affected by the UPS crash yesterday, are those all customers of your budget (low end) servers?

It's great if ezzi redundancy has improved much since last year and I may give it another try.


Yes we have greatly improved from last year, glad you have noticed :)

We do not separate any dedicated server customers.

We have built and manage separate data centers for larger corporations. Those are separate from anything we do with dedicated servers.

Originally posted by MrManager
EZZI, how would you define your 'core network'? I saw your explanation above, but that would seem to be your entire network... not just the core. Anyway, just curious.



Core equipment to me would be the equipment that routes all the traffic... I am sure everyone will have a different answer for this question.

Originally posted by ispclub.com
EZZI's core network has been up 100% this year. We may have had local issues here or there, like what happen yesterday that was very unfortunate and also out of our control... but the core network has been up 100% this year.


Except for about a half hour ago when the network was down for 30 minutes during a DoS. Packetloss was at 90% between ezzi-2.customer.tlw.nac.net and 65.125.239.146. Its nice to see the problem was resolved in 30 minutes on a Sunday, but what I expect from a company who advertises, "EZZI guarantees 99.9999% network uptime for Dedicated Internet Access customers." if you are going to have that many nines you have to be able to gracefully handle a DoS without an outage.


Problems will happen, it comes down to how fast the provider deals with it and also how they communicate with the customers letting them know what is going on.

The problem you referred to was pawned off as a UPS failure at the fault of the UPS maintenance people... this is the same reason that Hurricane Electric gave when they toasted most of their Freemont Facility. What is apparant however is that this excuse was not totally true. I'm a reseller with ezzi.net dedicated servers (means I've got 4 or more of them), and I had reboots/crashes happen the day before they said this problem occured and for four days afterwards. I had a total of 12 reboots happen, many of them AFTER the problem had been resolved.

When I opened up trouble tickets with ezzi.net about it asking them what is going on, their only response was, "Is it still happening?" No explanation, no apology, no offer of compensation for all the outages. Now this outage.

Maybe ezzi.net has improved, but it certainly isn't something you should put your mission critical equipment on that needs, as they describe, " EZZI guarantees 99.5% network uptime for dedicated server customers." UGH, I'm going elsewhere.

By saying that you had 100% uptime on your core network, this means you can still have a link or more go down, and your core network is still up, right? If that is the case, I am sure many DCs have 100% "Core Network Uptime", this does not account DoSs, saturated links, etc.

And, yet ANOTHER "one-post-wonder" miraculously appears here on WHT for the first time with a claim of horrible service to save us all from ourselves....

Originally posted by hacim
Except for about a half hour ago when the network was down for 30 minutes during a DoS. Packetloss was at 90% between ezzi-2.customer.tlw.nac.net and 65.125.239.146. Its nice to see the problem was resolved in 30 minutes on a Sunday, but what I expect from a company who advertises, "EZZI guarantees 99.9999% network uptime for Dedicated Internet Access customers." if you are going to have that many nines you have to be able to gracefully handle a DoS without an outage.
Funny... we haven't seen any packet loss, outage/downtime, or slowdowns as you have claimed since the UPS failure. And, our traffic goes through the 65.125.239.146 hop as well.
The problem you referred to was pawned off as a UPS failure at the fault of the UPS maintenance people... this is the same reason that Hurricane Electric gave when they toasted most of their Freemont Facility. What is apparant however is that this excuse was not totally true. I'm a reseller with ezzi.net dedicated servers (means I've got 4 or more of them), and I had reboots/crashes happen the day before they said this problem occured and for four days afterwards. I had a total of 12 reboots happen, many of them AFTER the problem had been resolved.No reboots/crashes here before or after the UPS failure as you claim. The UPS failure is the only outage we have seen in quite some time and there were no before issues or "after-shocks" that we have seen. To quote one of our servers, "System Uptime 7 Days, 11 Hours and 47 Minutes" which reflects it's been up since the UPS failure and not a second of downtime since then.
Maybe ezzi.net has improved, but it certainly isn't something you should put your mission critical equipment on that needs, as they describe, " EZZI guarantees 99.5% network uptime for dedicated server customers." UGH, I'm going elsewhere.
Ezzi has most definitely made some major improvements. Quite honestly, until you can back up your claims with proof and verify that you are even an Ezzi client (with 4 or more servers as you claim), I seriously doubt ANYTHING you say. I know what our experiences have been with them and it is NOTHING like you are claiming. What I DO know is that you apparently created a new account TODAY for the sole purpose of posting something negative about Ezzi:


hacim
Date Registered: 10-24-2004
Status: New Member
Total Posts: 1 (1 posts per day)
Last Post: 10-24-2004 05:39 PM
Ezzi Downtime History

Can you prove otherwise? How did it come about that your first post here is a slam on this company?

There was a DDOS attack yesterday ask ezzi your self.
I couldnt reach my servers.

Maybe you need another monitoring system.

Originally posted by Two-A-T
And, yet ANOTHER "one-post-wonder" miraculously appears here on WHT for the first time with a claim of horrible service to save us all from ourselves....


No, I came here looking for a different provider and saw this post and felt I needed to provide some of my experience. So what if this was my first post? Everyone has to post for the first time sometime.


Funny... we haven't seen any packet loss, outage/downtime, or slowdowns as you have claimed since the UPS failure. And, our traffic goes through the 65.125.239.146 hop as well.


Why dont you call up ezzi.net and ask them what happened yesterday, they will tell you with authority what I said about the DoS attack.


No reboots/crashes here before or after the UPS failure as you claim. The UPS failure is the only outage we have seen in quite some time and there were no before issues or "after-shocks" that we have seen. To quote one of our servers, "System Uptime 7 Days, 11 Hours and 47 Minutes" which reflects it's been up since the UPS failure and not a second of downtime since then.


To quote one of my servers:
root pts/1 adsl-58-73-85-13 Mon Oct 18 18:01 - crash (01:19)
root pts/0 adsl-58-73-85-13 Mon Oct 18 17:52 - crash (01:29)
root pts/1 adsl-58-73-85-13 Mon Oct 18 16:40 - crash (00:57)
root pts/0 adsl-58-73-85-13 Mon Oct 18 15:47 - crash (01:50)
root pts/0 adsl-58-73-85-13 Sun Oct 17 23:28 - crash (11:31)
root pts/4 adsl-58-73-85-13 Sat Oct 16 13:49 - crash (1+21:10)
root pts/1 adsl-58-73-85-13 Sat Oct 16 13:47 - crash (1+21:12)
root pts/3 sms1 Sat Oct 16 13:45 - crash (1+21:14)
root pts/2 adsl-65-73-85-13 Sat Oct 16 13:44 - crash (1+21:15)
root pts/1 adsl-65-73-85-13 Fri Oct 15 14:53 - crash (17:37)
root pts/2 sms1 Fri Oct 15 13:15 - crash (19:15)
root pts/0 adsl-65-73-85-13 Thu Oct 14 17:49 - crash (1+14:42)

These are managed servers, they shouldn't be crashing, the problem was on the 17th, so they claim, but you notice the crashes happened on the 14th, 15th, 16th, 17th and 18th. I was logged in on the 18th during every one of those crashes and I was on the phone with ezzi asking whats going on.


Ezzi has most definitely made some major improvements. Quite honestly, until you can back up your claims with proof and verify that you are even an Ezzi client (with 4 or more servers as you claim), I seriously doubt ANYTHING you say. I know what our experiences have been with them and it is NOTHING like you are claiming. What I DO know is that you apparently created a new account TODAY for the sole purpose of posting something negative about Ezzi:


How would you like me to prove that I am an ezzi customer? I don't care if you believe me or not. I believe that you didn't have the same crash/reboots that I've had, probably because your machines were in a different location. But just because you didn't notice the packetloss and your servers being completely unreachable for 30 minutes yesterday, doesn't mean it didn't happen.

You apologize for ezzi.net a little too much, makes me think you are an employee there.


Can you prove otherwise? How did it come about that your first post here is a slam on this company?

Dunno, I can't think of a way to prove that I am an ezzi customer... I could tell you something that is in the control panel that only customers can access? Or maybe I just don't care that much and don't feel the need to prove anything, believe me if you want, or don't, doesn't matter to me.

Why is my first post a slam of this company? Well, I had no other reason to post to these forums thus far. I found it particularly ironic that when I came here looking for an alternative I saw all these rosy posts about ezzi, so I felt I needed to share my experiences with others.

Hacim,

Stop accusing others of being employees of EZZI, just because they don't appear to have the same problems you have.

So what if there are a lot of good reviews about EZZI? No one here is apologising for anyone.

Are you going to accuse me of being an employee as well?

Originally posted by Darkanoid
There was a DDOS attack yesterday ask ezzi your self.
I couldnt reach my servers.

Maybe you need another monitoring system.
If there was in fact a dDoS attack yesterday, that does not mean it affected every server in the DC. It could have been isolated to one rack, one router, etc. We never saw it and there were no reports from clients about problems with the exception of one client who is having occasional DNS routing issues at his ISP which is not Ezzi's problem and another in Spain who is having a similar issue.

I do not base our outage reports only on our own monitoring. We have quite a few clients who have mission critical sites and will contact us if there is even a "burp" in the network that affects them.

We had 3 servers affected yesterday (different IP ranges as well), and infact the ezzi.net was also not available during the attack so we could't even find out what was going on at the time, luckily it didn't last long.

Saying that however we have been very happy with ezzi so far, response time's have been fairly good (under an hour for most things, longest wait was about 6 hours for a minor question). Although we did order another server Friday night and it's still not up which is a lot longer than we have waited for other servers in the past.

Originally posted by poolking
Hacim,

Stop accusing others of being employees of EZZI, just because they don't appear to have the same problems you have.



So let me get the rules straight: I can be accused as not being a customer if I am critical, but I am not allowed to accuse someone as an employee? This sounds like the ole "dish it out, but can't take it" argument.


Are you going to accuse me of being an employee as well?

No, I'm not sure why I would... you aren't accusing me of faking being a customer and you aren't posting things that are too positive of the company in an effort to try and cover up some criticism.

Originally posted by hacim
No, I came here looking for a different provider and saw this post and felt I needed to provide some of my experience. So what if this was my first post? Everyone has to post for the first time sometime.
Maybe you are legitimate... but the fact remains that there have been quite a few cases of similar attacks that were proven to be false so when a new member's first post here is an attack on a provider, it immediately throws up red flags.
Why dont you call up ezzi.net and ask them what happened yesterday, they will tell you with authority what I said about the DoS attack.
There may have been an attack. If there was, we never saw any signs of it ourselves nor did any of our clients report issues or outages.
These are managed servers, they shouldn't be crashing, the problem was on the 17th, so they claim, but you notice the crashes happened on the 14th, 15th, 16th, 17th and 18th. I was logged in on the 18th during every one of those crashes and I was on the phone with ezzi asking whats going on.
Ok.. as far as this goes, managed or not, a server can be brought down by something as simple as a single line of bad code in one of the sites so this means nothing. Also, individual server crashes are not necessarily Ezzi's fault. One server (or even several) does not constitute an outage or provider failure. It simply tells me there could be something on your servers that is causing a crash. If one of our servers has repeated crashes, I look at the server itself and not the DC.
How would you like me to prove that I am an ezzi customer? I don't care if you believe me or not. I believe that you didn't have the same crash/reboots that I've had, probably because your machines were in a different location. But just because you didn't notice the packetloss and your servers being completely unreachable for 30 minutes yesterday, doesn't mean it didn't happen.
And the apposite is also true. That's the key here.. just because you DID have problems doesn't mean it was affecting the entire DC and therefore it does not mean Ezzi is a horrible provider.
You apologize for ezzi.net a little too much, makes me think you are an employee there.
No, it just means I have gotten tired of the attacks against them. I have been dealing with Ezzi for close to 2 years now and with the exception of some major issues they had at the end of last year, I have not seen most of the things that some like to attack Ezzi for. And given the fact that in most cases the attackers fail to provide proof of their claims (and some even suddonly disapear when others like myself ask for proof), I have a hard time believing them.

Do not mistake my defense of a company I believe to be a quality provider based on our own experiences as apologies. I am the type of person who will only make negative claims against someone when I can back up those claims with proof. I also believe in loyalty. This means when a provider treats me well and does their best to provide the services I pay them for, I will back them up when others attack them. Ezzi has never failed to provide us with fast, professional and quality services even during the major problems they had last year. We were always given credit when outages warranted and we have always been taken care of. So, while not impossible, I do find reports of the opposite hard to believe and I will therefore challenge anyone who makes such claims without providing proof and anyone that registers new here to do it. Would you like it if one of your customers believed negative claims about your company without challenging them when the claims are opposite of their own experiences? Or, would you prefer they respond with their own positive experiences to counter the negative (and possible false) claims?

By the way.. if I WERE an employee, why would I challenge you being legitimate? I would know that you are a customer and I would know that there was an outage.
Dunno, I can't think of a way to prove that I am an ezzi customer... I could tell you something that is in the control panel that only customers can access? Or maybe I just don't care that much and don't feel the need to prove anything, believe me if you want, or don't, doesn't matter to me.
Unless ispclub challenges your claims of being a client, I will accept that you may in fact be one.
Why is my first post a slam of this company? Well, I had no other reason to post to these forums thus far. I found it particularly ironic that when I came here looking for an alternative I saw all these rosy posts about ezzi, so I felt I needed to share my experiences with others.
As I said... any time someone's first post is an attack on a provider, it immediately throws up red flags and I have doubts as to the validity of their claims. I'm sure you would want others to think the same if it happened to you.

Originally posted by hacim
So let me get the rules straight: I can be accused as not being a customer if I am critical, but I am not allowed to accuse someone as an employee? This sounds like the ole "dish it out, but can't take it" argument.



No, I'm not sure why I would... you aren't accusing me of faking being a customer and you aren't posting things that are too positive of the company in an effort to try and cover up some criticism.

So how can someone be an employee of one company, but have another webhosting company in their signature, does that make any sense to you?

He was not covering up any criticism he was posting his own experiences. It is just that you cannot except that people have positive things to say about EZZI.

Originally posted by hacim
So let me get the rules straight: I can be accused as not being a customer if I am critical, but I am not allowed to accuse someone as an employee? This sounds like the ole "dish it out, but can't take it" argument.
No, the concern that you may not be a ligetimate customer was made because your first post in these forums was an attack on a provider, your profile has no information in it to indicate that you are legitimate (such as website URL, screen names, etc) and your claims were different from our own experiences. These things can easily be warning signs of a false claim. But, if you are a regular here, have details about yourself in your profile and provide proof of your claims that's a different story. Yes, first time posters who attack a provider can be legitimate. But, without any way to verify their claims, there can easily be doubt. Stick around these forums for a while and pay attention to the "one post wonders" who come along and you will see what I mean. Let's hope YOU are never the target of one of them... given your statements here, you would have no grounds to challenge them even if the claims were hard to believe.
No, I'm not sure why I would... you aren't accusing me of faking being a customer and you aren't posting things that are too positive of the company in an effort to try and cover up some criticism.
How have I tried to cover anything up? I merely stated that what you are claiming, we have not seen. And, yes I post positive about Ezzi because our experiences with them have been very positive. I have no doubts you would want the same from your own clients if their experiences were positive and different from those of a first time poster who only had negative to say.

Well I decided to let things drop because I don't feel like arguing on some webboard. I dont really think you are an employee, I just found it odd that you would accuse me of deliberately trying to sabotage ezzi.net, but I understand your reasons now.

However... I am writing this while I am on hold with ezzi.net. Six of our machines are not available again. They went to check on something, and suddenly they came back, I logged into them quickly, one had been rebooted (it had only been up one minute), one had been rebooted 20 hours ago, the others were still on. While I was on them, I lost connectivity again. A few minutes later I am able to get to them again. This time 4 of them were rebooted, two of them were not rebooted. WHY ARE THEY REBOOTING MY MACHINES?! Obviously this has nothing to do with the 58% packetloss I am getting, and rebooting them without talking to me certainly hasn't helped. I've been managing linux servers for 13 years, I know when I see wrecklessness when I see it. Either they've got someone who doesn't know what they are doing and think rebooting fixes things, or they've got some weird power problem. I'm still on hold 15 minutes later... considering calling back to ask WTF?!

So I waited a few more minutes, nothing... then I hung up and called back. I started to explain that I was on hold with someone and they never came back and then mid-sentence I was interrupted, "Hold on a sec!", a couple seconds later I was in a voicemailbox... so I called back, nobody answers. Gah.

Originally posted by hacim
Well I decided to let things drop because I don't feel like arguing on some webboard. I dont really think you are an employee, I just found it odd that you would accuse me of deliberately trying to sabotage ezzi.net, but I understand your reasons now.

However... I am writing this while I am on hold with ezzi.net. Six of our machines are not available again. They went to check on something, and suddenly they came back, I logged into them quickly, one had been rebooted (it had only been up one minute), one had been rebooted 20 hours ago, the others were still on. While I was on them, I lost connectivity again. A few minutes later I am able to get to them again. This time 4 of them were rebooted, two of them were not rebooted. WHY ARE THEY REBOOTING MY MACHINES?! Obviously this has nothing to do with the 58% packetloss I am getting, and rebooting them without talking to me certainly hasn't helped. I've been managing linux servers for 13 years, I know when I see wrecklessness when I see it. Either they've got someone who doesn't know what they are doing and think rebooting fixes things, or they've got some weird power problem. I'm still on hold 15 minutes later... considering calling back to ask WTF?!

Well I decided to let things drop because I don't feel like arguing on some webboard

I'm still on hold 15 minutes later... considering calling back to ask WTF?!

So you have been working on linux servers for over 13 years. This means you have never seen webhostingtalk? You sound condescending on WHT. ...

So your like min. ~28 years old, and you go around saying "WTF?!" ?

You sure about that?

Something doesnt add up for me.

Got them on the phone again... turns out it was a bad UPS, again. They say they got a bad "batch" of them. I suppose that could've made one switch go out, and not another, but boy... I certainly would be looking for another UPS manufacturer, preferably an enterprise reliable one who you have contracts with so when things like this happen. I am sure that they want to provide the good service that they advertise and are incredibly disappointed about this, I'm not going to say that they are malicious or anything. However, I've worked in setting up telco grade datacenters, and you don't say 5 nines like they do unless you've got redundant power. That doesn't mean a UPS for when the power goes out, that means dual grids if possible, carrier grade UPS with diesel generator backup. I personally don't think they should advertise 99.99999% uptime because that is out of their league.

We've had two servers suffer quite badly from the "UPS problems", however the errors we're seeing point to a faulty NIC, although the techs are obviously more experienced than I am at finding problems. I reported both server problems again tonight as was told it was a faulty UPS effecting both servers, I will continue to monitor the situation closely as both users are getting rather annoyed with the constant problems.

Originally posted by hacim
Got them on the phone again... turns out it was a bad UPS, again. They say they got a bad "batch" of them. I suppose that could've made one switch go out, and not another, but boy... I certainly would be looking for another UPS manufacturer, preferably an enterprise reliable one who you have contracts with so when things like this happen. I am sure that they want to provide the good service that they advertise and are incredibly disappointed about this, I'm not going to say that they are malicious or anything. However, I've worked in setting up telco grade datacenters, and you don't say 5 nines like they do unless you've got redundant power. That doesn't mean a UPS for when the power goes out, that means dual grids if possible, carrier grade UPS with diesel generator backup.
I have physically been to the Long Island DC myself and have seen their backup systems. They do have several levels of UPS and redundant power (with more power being added which is what caused the UPS failure last week when one of the electricians caused a surge). What probably caused your situation was the smaller UPS's they have on each rack which would be the last point of power on your rack before reaching the servers. If that UPS fails, there is no way to provide secondary power because your server is actually powered from that UPS.

If memory serves me correctly (I'm sure Phil will correct me if I am wrong or if it has changed since I visited the DC almost a year ago now), there are at least 2 levels of UPS protection in the DC: Larger industrial UPS's that feed power to smaller rack-level UPS's which provide power to the servers. In this case, if the smaller rack-level UPS fails, that breaks the circuit your server (and any other on that rack) is on and there's nothing that can be done to guarantee power.

If you have been involved in built-outs like you say then you should understand that there will ALWAYS be that last redundant or backup point at which if a failure occurs there will be a power loss or fluctuations and there is no way to stop it.

I personally don't think they should advertise 99.99999% uptime because that is out of their league.
They don't guarantee the 99.9999% uptime for dedicated servers, they guarantee that uptime for their dedicated internet access clients:
EZZI guarantees 99.9999% network uptime for
Dedicated Internet Access customers.

EZZI guarantees 99.5% network uptime for dedicated
server customers.

Wow, everyone is vicious around here. Scares me!

Originally posted by Two-A-T
Larger industrial UPS's that feed power to smaller rack-level UPS's which provide power to the servers.
Let me get this straight... they've invested heavily in datacenter grade system UPSes for their facility, which should be fault tolerant and extremely reliable; and they've then hung low-end rack mountable UPSes meant to power a few servers, or perhaps a rack worth of them, off their system UPS?

That's a pretty stupid design. The large investment in a system UPS buys you redundancy on things from electronics that control the UPS to bypass operation that can feed your load directly off the grid so you can perform battery maintenance, etc. Low end UPSes meant to feed a rack worth of servers do not have these features, and consequentially, do not improve load uptime; indeed, they reduce it substantially.

I suspect that Ezzi isn't dumb enough to miss something obvious like this, and that you are just unfamilar with the design of their power system.

If you have been involved in built-outs like you say then you should understand that there will ALWAYS be that last redundant or backup point at which if a failure occurs there will be a power loss or fluctuations and there is no way to stop it.
Really? In telco, we powered most things from redundant -48VDC buses. This meant that no part of the power plant was a single point of failure that could result in a load being de-energized unintentionally. Today, I do this with many pieces of network infrastructure (routers, core switches, etc.) as an additional safety measure, whether or not the devices offer redundant AC power.


They don't guarantee the 99.9999% uptime for dedicated servers, they guarantee that uptime for their dedicated internet access clients:
That's correct, but who knows what they consider &quot;up&quot; with those kind of numbers. Note that their guarantee allows for < 32 seconds of downtime per year for Internet access; and 3.65 hours per month of downtime for dedicated server customers. That doesn't include hardware or software failures on your server, mind you; that's network uptime, which is the easiest number to get.

I suspect their customers don't really experience downtime even close to breaching that 99.5% guarantee, but if they do, man, you'd have to be amazingly cheap to remain an Ezzi customer.

Don't just see 99.5% network availability, or six nines of leased line availability (why is it they can provide good uptime over telco circuits they don't control, but can't provide good uptime in a datacenter network they have 100% control over, by the way? go figure.)

Get out your calculator, figure out how much downtime that really allows for, and ask yourself if you are willing to accept that kind of downtime if it happens. My clients would fire me in a second if I couldn't engineer their networks for better than 99.5% availability; and their customers would be jumping ship to other vendors before you could count 3.65 hours down.