Anyone else's Eumx gone screwy? - Page 3

17pm · 10 Oct 2016, 11:26 PM

Quote:

Originally Posted by curtiscarmack

I can see now that I am on the new platform. When I signed up, I was told that it would be the main platform within 30-60 days (that was well over 10 months ago). I spent the first 2 months testing, and all seemed to be well. My mistake for assuming it was out of testing.

Damn, that sucks.. I guess this cements the idea that EuMX is not playing in the big league (primary e-mail services).

Berenburger · 11 Oct 2016, 08:07 AM

Quote:

Originally Posted by 17pm

Damn, that sucks.. I guess this cements the idea that EuMX is not playing in the big league (primary e-mail services).

That's right.

curtiscarmack · 12 Oct 2016, 06:53 AM

Well, it's been a week now since my last access to email and no replies from EuMX on when, if ever, I will get access again. If anyone has heard anything, please post, as I am losing faith that I will ever hear back through a normal support channel.

fabule · 12 Oct 2016, 06:22 PM

Well,

This October is clearly not our month. Storage failures, database corruption, btrfs* file system corruption on new infrastructure and now total network outage.

Replica systems in Germany works, but they are not accessible for public at this moment. We hope it can't go even worse and from this point we will quickly recover our service and it's reliability you expect from us.

We will replace btrfs and go with traditional fs to production.

Quote:

Originally Posted by curtiscarmack

I can see now that I am on the new platform. When I signed up, I was told that it would be the main platform within 30-60 days (that was well over 10 months ago). I spent the first 2 months testing, and all seemed to be well. My mistake for assuming it was out of testing.

fabule · 12 Oct 2016, 07:00 PM

Service status:

- Network issue with peering partner is fixed,
- Storage is recovered,
- Database cluster is recovered,
- All file systems are recovered,
- btrfs is replaced to proven reliable one,
- new infrastructure servers are synchronized with replicas,
- all mails are delivered from queues,

- Redundancy is still not fully restored. Work in progress.

curtiscarmack · 13 Oct 2016, 08:05 AM

I'm happy to report that I have access again to email on the new infrastructure. From what I can see, emails from the first 30 or so hours of the outage have not been recovered, but perhaps they will yet show up. It may be that they were undeliverable for too long and failed. Happily the rest of the week's worth of email showed up this morning.

Perhaps the EuMX folks will eventually get some rest -- I'm sure they need it.

randian · 15 Oct 2016, 06:07 PM

Quote:

Originally Posted by fabule

We will replace btrfs and go with traditional fs to production.

Why not ZFS? Same benefits, proven reliability.

12 Oct 2016, 06:53 AM	#33
curtiscarmack Junior Member Join Date: Oct 2016 Location: Lehi, Utah, USA Posts: 7	Sadly no replies from EuMX Well, it's been a week now since my last access to email and no replies from EuMX on when, if ever, I will get access again. If anyone has heard anything, please post, as I am losing faith that I will ever hear back through a normal support channel.

13 Oct 2016, 08:05 AM	#36
curtiscarmack Junior Member Join Date: Oct 2016 Location: Lehi, Utah, USA Posts: 7	Access is back for me I'm happy to report that I have access again to email on the new infrastructure. From what I can see, emails from the first 30 or so hours of the outage have not been recovered, but perhaps they will yet show up. It may be that they were undeliverable for too long and failed. Happily the rest of the week's worth of email showed up this morning. Perhaps the EuMX folks will eventually get some rest -- I'm sure they need it.

12 Oct 2016, 07:00 PM	#35
fabule Essential Contributor Join Date: Jan 2005 Posts: 413 Representative of: eumx.net	Service status: - Network issue with peering partner is fixed, - Storage is recovered, - Database cluster is recovered, - All file systems are recovered, - btrfs is replaced to proven reliable one, - new infrastructure servers are synchronized with replicas, - all mails are delivered from queues, - Redundancy is still not fully restored. Work in progress.