nikolas
@lightpath.net
|  Earthlink Out(r)age
The fact that a major (soon to be minor) provider has neither a reliable power generators nor a failover plan does not bode well for their future, to say the least.
Incidentally, this is third time in two years that Earthlink-generated outage causes a major problem for my business. We will be Earthlink-free ASAP. |
|
cdru
Go Colts
Premium,MVM
join:2003-05-14
Fort Wayne, IN
| said by nikolas :
The fact that a major (soon to be minor) provider has neither a reliable power generators nor a failover plan does not bode well for their future, to say the least. Earthlink isn't the first nor the last to find out that a backup generator isn't generating quite like how they wanted. Usually companies find out these types of things only when their service is call upon and they don't work. Unfortunate, yes. End of the world, no. SMTP is resilient. It will survive a few hours or even a day or two with downtime. The summary doesn't even state that the problem is necessarily within Earthlink's direct control.
Incidentally, this is third time in two years that Earthlink-generated outage causes a major problem for my business. We will be Earthlink-free ASAP. Well, why don't you have a failover plan? It's not hard to have a backup mail server. You'd think you would have had a failover plan after the first major problem for your business... |
|
nikolastojsi
join:2003-07-30
New York, NY
| We are actually very good failoever-wise, and we do have multiple email servers - on two separate ISP links, and with a TESTED backup generator, in fact. However, Earthlink is still hosting our outside DNS (mainly because our outside domain was purchased through them - long ago and a long story), and they refuse to change the SOA record. That mess requires 72 hours to straighten out (according to Earthlink), so it is scheduled to change during the next long weekend (7/4).
IMHO, it is not end of the world, by no means - but it is the end of Earthlink, also IMHO.
Btw, I am happy that you are holding us to higher standard than you do Earthlink! |
|
exocet_cm
In memory of dadkins
Premium
join:2003-03-23
New Orleans, LA
clubs:   | reply to cdru
Even I have TrippLite 2 and 6 hour UPS on all of my home APs, hardware firewall, gaming/dvr. |
|
k1ll3rdr4g0n
join:2005-03-19
Homer Glen, IL
| reply to nikolastojsi
said by nikolastojsi  :We are actually very good failoever-wise, and we do have multiple email servers - on two separate ISP links, and with a TESTED backup generator, in fact. However, Earthlink is still hosting our outside DNS (mainly because our outside domain was purchased through them - long ago and a long story), and they refuse to change the SOA record. That mess requires 72 hours to straighten out (according to Earthlink), so it is scheduled to change during the next long weekend (7/4). IMHO, it is not end of the world, by no means - but it is the end of Earthlink, also IMHO. Btw, I am happy that you are holding us to higher standard than you do Earthlink! Why did Earthlink even manage the DNS in the first place? You should be able to change DNS servers. If you couldn't that should have been a red flag to grab your auth code, unlock the domain and move to a different registrar. |
|
battleop
join:2005-09-28
00000 | reply to nikolas
You can have all the redundancy in the world and still have an outage. It can happen to anyone big or small. Nothing is 100% perfect. |
|
webnetwiz
There's no place like 127.0.0.1
Premium
join:2004-09-22
Van Nuys, CA
| reply to nikolas
Dude, you don't know what you're talking about. I used to work at the Pasadena data center, and believe me, it has plenty of generators in the back, and they have very good procedures for testing and such. This is more likely system related than an actual power issue. A lot of times companies blame systems issues on "power", because its easier to explain to customers. And they do operate two hubs that back each other up, one in PAsadena and one in Atlanta. This is got to be some system screw up, I remember they used to have them all the time especially with call center apps. |
|
bac522
join:2003-08-04
Manchester, NH
| Dude...you don't know what you are talking about! No matter how much you think you have redundancy built in or how many remote sites you think are reliable there is always a SPoF (Single Point of Failure) you didn't think of. The AT&T fiber cut a few weeks back was a prime example of that! |
|
nikolas
@rcn.com
| reply to k1ll3rdr4g0n
said by k1ll3rdr4g0n : Why did Earthlink even manage the DNS in the first place? I happen to agree with you that that is not a good idea - but one can make an argument that it makes sense to have a major provider manage outside DNS (which should result in a faster DNS response time etc. etc.). And I cannot really blame people who made that decision - it was done many years ago, in the dialup era (the company got domain registration and hosting, and web and email hosting AND a couple of dialup accounts - the most important part of the deal at the time! - through Mindspring for a reasonable price).
said by k1ll3rdr4g0n : You should be able to change DNS servers. If you couldn't that should have been a red flag to grab your auth code, unlock the domain and move to a different registrar. Once again, I could not agree more, we should be able. Emphasis on SHOULD. And it was a red flag. I had immediate discussions with Earthlink and Network Solutions (our new registrar of choice) how to proceed with the domain SOA and DNS move. The tech from Network Solutions happened to have previous experience with the company that was in the same situation as us - domain registered through Earthlink, with underlying ownership uncertainties - and, according to him, their outage was 72 hours. (Having spoken with Earhlink techs at the time, I believed it - they were less than willing to facilitate SOA transfer, to say the least.) Now it looks like we will have to do it on an emergency basis - I really do not see Earthlink staying in business much longer.
Look, this thread is really about Earthlink - not about my company. (I feel we are hijacking the thread.) Whatever we did or did not do is IMHO beside the point - the fact that Earthlink, an ISP and service provider, did not have (1) adequate backup generators and/or (2) at least one failover location - and thus allowed itself to have nationwide outage lasting six hours is a HORRIBLE statement about their business. |
|
