Mail server hosed

Mail server ( appears to be hosed. Sending N emails to a DreamHost email address and observing the packet traces via tcpdump shows that messages to mx1 go through while mail to mx2 goes into a black hole (with only a subset of messages turning up, often hours later).

I’ve submitted several support requests to DreamHost support – no reply whatsoever. Has anyone received a response from them regarding this?


how long did you wait to hear back from DH? You may have to be patient. but they will get back to you. It could be that they are actually working on the problem as we post, and are doing that rather then not-working on the problem to E-mail you to say they are trying to work on the problem…


More specifically, I shouldn’t say go into a black hole. The messages are received in both cases by mx1 and mx2. It’s just that the messages that go through mx1 are actually received at the destination’s email address. In contrast, messages that go through mx2 are either delayed or dropped on the floor. Example: sending 100 emails with subjects 1,2,3,…,100 to results in only a subset actually reaching and this subset is precisely the set of messages that go through mx1.

Can someone at DreamHost just turn off mx2 until you fix this problem? Sure, the mx records will take a while to propagate but mail servers should be smart enough to not try load balancing across mx2, which is perpetually hosed apparently.


The first message was sent at around 8:25am PST today (4/18). Subsequent messages were spaced out several hours apart.


well, there’s a couple of possibilties for support.

First and fore most, they say that support will be responded to within 24 hours. So do give them a bit of time. Also, did you set your request as a high priority?

And you really shouldn’t send in so many support requests, its just not nice.


Well, this has been going on for almost two days now. I gave them a day to work it out before I even sent in a support request since presumably they were aware of the problem since this affects all mail going through that machine. And yes, the request was set at high priority. I guess I’ll just wait around for another 12 hours while my mail gets arbitrarily delayed/dropped since setting a request as high priority obviously doesn’t have any impact on priority, except perhaps once the request has been in the queue for, as you mention, 24 hours.

Yeah, that was a mistake. I will definitely make sure to whine to them more frequently in the future the minute I notice something is troubled on their end – or, better yet, write an external monitor of their critical services and send them automated emails via my own account when things start to melt down.


PS This mail problem still hasn’t been fixed as of 1:33am PST the next day (right now). I just sent myself 20 emails a while ago and received 12/20 immediately, i.e., the 12 messages that used mx1, rather than mx2. I’m still waiting for the remaining ones, some subset of which will probably trickle in over the next few hours.

A customer was flooding that mx server and delaying mail for other users. We did not notice it right away, as you pointed out.

Sending us automated emails will most likely just give you a bad reputation with our support team. If that is your intention then feel free to do so, but if you really just want to help us improve our service there are better ways to do so.

Please do notify us if you believe there is a problem on our end. Setting a support message to high priority does get noticed, but remember that some problems just take time to resolve. We do not typically respond to messages until the problem is resolved so a long wait for a response does not mean we have not seen the message at all.

This problem should have been resolved much quicker than it was and I do apologize for that. We may have to start restricting the use of our mail servers more to prevent issues like this from happening.

  • Dallas
  • DreamHost Honcho