Mail10 6/10/2008

[ericfoy]

I sincerely hope this works! I've struggled with mail10 and its ill performance for months (like, at least twelve of them) now. This thing has always been flaky. It'll be nice to have reliable mail service for a change. No second chances here, guys, so take your time tonight and get it right!

[JoshK]

New machine is up and running, setting up new drives for it. Expect it to be online in a couple of hours. I will post up the second it's back in the racks!

[micn]

It is very sad that all mail either outgoing or incoming are down for all our users! Now, we cannot even login to the cpanel!!

[micn]

I notice we are designated as Junior member, umm, I wonder what makes a Senior member!!

[JoshK]

It is based purely on forum activity not on longevity as a customer! Please dont' take offense to it!

[melbinfotech]

Server is up !
POP3 is working fine.
Cheers~

[JoshK]

Well it Looks like I was beaten to it! Yes it's back up, new machine seems to be working quite well, already caught back up, load nice and low. We have not enabled spamd yet, we're going to go one step at a time, adding functions to make sure it's completely stable at each step, but all standard mail functions should be back to normal now.

[dvanburen]

It's been up for about 30 minutes. So far so good, but I said that last time too...

It's on completely new hardware. If the same problem recurs we will have to reimage the box and transfer over the config files, and possibly rebuild the data drive (read: BAD). Hopefully this won't be necessary.

[kapuwa]

Did we loose all old emails?

[ekrubonline]

Still getting complaints. I submitted a ticket about it. My clients are sending average 10 emails per message - and that's when they can actually send emails.

[dvanburen]

Originally Posted by ekrubonline

Still getting complaints. I submitted a ticket about it. My clients are sending average 10 emails per message - and that's when they can actually send emails.

That's the only complaint we have had so far, and we cannot reproduce it. The load is normal and there are only a nominal amount of connections to the server. Please make sure you are using port 2525 to send mail. If your smtp sessions time out you may want to try a ping/trace route to the server to see where the bottleneck is.

[JoshK]

you did not loose any data, all data was recovered from the old machine and moved to the new machine.

REMEMBER EMAIL IS NOT BACKED UP! It is considered volatile storage. Old messages are removed at some point when they reach 30-45 days old.

[Silverbug]

Once you guys get it sorted can you give us a run down on how long each step took. My thinking is if something similar happens again people can start to see from past rebuilds and such, how long they should wait before jumping up and down.

just my 2c

[JoshK]

it depends in each case, for instance, everytime the server was overloaded due to spam attempts it took anywhere from 15 minutes to 1 hour to reboot, clear it, lock it down and be back to going again. Software copies take from 2 hours to 10 hours depending on size of drives involved and number of files. That of course is also depending on drive condition, a failing hard drive can bring that time figure into the 2 or three 3 range. It never hurts to tell us when you see a problem, It is often impossible in the early stages to give an estimate on repair time as often it is the diagnosis that take the most time. Catch 22.. if we knew how long it would take, it would mean we already know what was wrong fully, and it would already be repaired. In this case we spent time copying the drives to new drives because it appeared to be a drive wearing down (considering it was being hammered by spaming and full activity for days leading up to the failure, that would be a logical path) but all of that was futile, as we had deeper hardware issues. Each repair strategy is different and hence the times involved can vary dramatically. Rest assured we NEVER want a repair to go any longer than absolutely required!

[Brangwyn]

Thanks for your efforts Josh.