Slashdot Mirror

Simon Anderson Says:
January 21st, 2012 at 11:55 am

some more detail – our systems have stored and used encrypted passwords for a number of years, however the hacker found a legacy pool of unencrypted FTP/shell passwords in a database table that we had not previously deleted. We’ve now confirmed that there are no more legacy unencrypted passwords in our systems. And we’re investigating further measures to ensure security of passwords including when a customer requests their password by email (this was not the issue here, though). Re your shell accounts, I’d suggest that you select a new password just to be sure.

Search for "January 21st, 2012 at 11:55 am" at this link

Also, due to the number of customers changing their passwords, the password sync time is very slow right now. More info here.

I'd be hard pressed to believe you've dealt with anyone other than DreamHost. When I failed to renew my service with them it was because their hosting was glacial. It could barely keep up with a lightly used PHP image gallery. That was years ago. When I migrated clients away from DH last year it was because of chronic downtime. "Oops we fucked up" is great, and it's honest. It's also not something you want to keep seeing. "Oops we fucked up, but your worthless blogs about your kittens' trousers are safe" is /not/ something you want to see, ever. We're suffering a DDoS... no... wait... we don't know how to fix our Cisco equipment... is not something you want to see ever. Certainly it's not something you want to keep seeing

Not doing any manner of scheduling for intrusive maintenance is not simply a tactic to scare away high maintenance customers, it's a tactic to scare away paying customers. You wanna know what's even less professional? Not having any phone support in the first place, and then not having any e-mail support or publicly available system status because everything's on the same network. A single point of failure isn't a tactic to "scare off the clueless, high-maintenance market" it's a hallmark of someone who doesn't know what they're doing.

I'd be hard pressed to believe you've dealt with anyone other than DreamHost. When I failed to renew my service with them it was because their hosting was glacial. It could barely keep up with a lightly used PHP image gallery. That was years ago. When I migrated clients away from DH last year it was because of chronic downtime. "Oops we fucked up" is great, and it's honest. It's also not something you want to keep seeing. "Oops we fucked up, but your worthless blogs about your kittens' trousers are safe" is /not/ something you want to see, ever. We're suffering a DDoS... no... wait... we don't know how to fix our Cisco equipment... is not something you want to see ever. Certainly it's not something you want to keep seeing

Not doing any manner of scheduling for intrusive maintenance is not simply a tactic to scare away high maintenance customers, it's a tactic to scare away paying customers. You wanna know what's even less professional? Not having any phone support in the first place, and then not having any e-mail support or publicly available system status because everything's on the same network. A single point of failure isn't a tactic to "scare off the clueless, high-maintenance market" it's a hallmark of someone who doesn't know what they're doing.

I'd be hard pressed to believe you've dealt with anyone other than DreamHost. When I failed to renew my service with them it was because their hosting was glacial. It could barely keep up with a lightly used PHP image gallery. That was years ago. When I migrated clients away from DH last year it was because of chronic downtime. "Oops we fucked up" is great, and it's honest. It's also not something you want to keep seeing. "Oops we fucked up, but your worthless blogs about your kittens' trousers are safe" is /not/ something you want to see, ever. We're suffering a DDoS... no... wait... we don't know how to fix our Cisco equipment... is not something you want to see ever. Certainly it's not something you want to keep seeing

Not doing any manner of scheduling for intrusive maintenance is not simply a tactic to scare away high maintenance customers, it's a tactic to scare away paying customers. You wanna know what's even less professional? Not having any phone support in the first place, and then not having any e-mail support or publicly available system status because everything's on the same network. A single point of failure isn't a tactic to "scare off the clueless, high-maintenance market" it's a hallmark of someone who doesn't know what they're doing.

We've been very pleased with Dreamhost -- they seem to be very friendly and security-conscious. (Note that they're also using Ksplice, which we've covered before, to keep their systems up to date without rebooting.)

My host is dreamhost. Their uptime is also listed as 99.99% ( http://www.webhostingstuff.com/uptime/DreamHost.html ).

From experience, this is not the case. For example, the site says that they had 100% uptime in January of this year. On their own status page ( http://www.dreamhoststatus.com/ ) they show several outages for January. There was one night in January where some DH routers had issues and the Web servers and MySQL servers could not communicate. Any web site which did not require MySQL was fine, but those that did require it were not working.

There was also a couple DNS outages late last year, in which all websites that used DH's DNS were down (web servers were fine, its the DNS servers that were not).

I'm not complaining about DH, as they are fantastic compared to 1and1, who I was with before. I'm just saying that I don't believe 99.99% uptime is accurate for DH (it may be accurate for Yahoo!), and 100% for the past 6 months is incorrect.

Also looking at the site for 1and1 ( http://www.webhostingstuff.com/uptime/11InternetInc.html ), I know those numbers are not correct. I left 1and1 after my site had daily outages in March/April 07.

Storage space has been a big issue of contention on Dreamhost as well. I signed up for their service, feeling happy that I had 500G of remote storage to use as I pleased. It turned out it wasn't that simple.

Unlimited sounds great, until you start using a large amount of space and Yahoo has to find some reason to say that you're not complying with their terms of service.

Basically someone put future dates in the billing system, making it believe we were in a future date, and resulting in ridiculous bills being sent out to every customer for a total of $7,500,000 in the short period of time the program run.

More info on the dreamhost page: http://www.dreamhoststatus.com/
And explanation of what happened by the guy who did it on its blog: http://blog.dreamhost.com/

http://www.dreamhoststatus.com/2007/11/17/gmail-forwarding-slowness/#comments

I'm posting as DS. (Which is what my slashdot username is, in binary).

A ton of people are up in arms blaming dreamhost. It seems that 90% of the people there are forwarding all their e-mail to gmail. I don't know why they don't use their own client or Google Apps. And they're complaining to HD that it's entirely their fault and they don't understand why it happened in the first place.

I was doing a catchall->gmail->myuser. This way gmail filtered all my spam but I also got to use dreamhost. But given this current issue I'm going to try out Google Apps.

I thought it was all or none. I guess you can set up JUST e-mail (I'm using calendar too though). You have to change your MX servers at Dreamhost. Coupled with IMAP access, this should work for me.

They aren't very redundant; see:

http://dreamhoststatus.com/

for examples. They have dozens/hundreds of shared servers and a pile of other infrastructure, so the status page gives the impression they are down a lot. On the other hand, I haven't actually ever noticed my site being down, even when I see a notice that something has gone wrong.

Dreamhost (my hosting provider) is having the same problem. Check out the excellent summary of the situation in this blog entry.

I have a feeling that it's a lot more than just two ISPs.