Safe-mail.net is down for critical maintainance (maintenance)
- Thread starter EQ Admin
- Start date
They have migrated 80% of users to a new storage system prior to this crash (reason of the migration could be speculated as the old hardware was failing). They are restoring the other 20% from the failed storage, or more likely, backups. They also seem to be taking this opportunity to clean up users that violated the ToS. This could speed up the restore if those accounts are simply ignored during the restore. They are going to put free users on a separate system and put premium users on a more HA system, to dampen impact on paying users for any further outages. Free accounts will no longer be able to be opened automatically. There is likely no way of knowing at this time if you were already migrated or not. If you were in the unlucky 20%, I would wager to say you've lost all of you email in the past month at least.
This is pure speculation. I'll take a shot at it.
It's strange to me that they mixed a storage issue with accounts that violate their TOS. It makes sense that they are trying to speed up their recovery time by restoring all premium accounts and free accounts that are active. It would make sense if they didn't restore, or restored last, the inactive accounts. Maybe their design makes that harder to do than it's worth.
They stated accounts that violate their agreement and policy will not be restored, not inactive accounts, which leads me to think it's spammers and/or accounts identified to be involved in illegal activity that will not be restored.
I suspect that Safe-Mail is a "medium" size email service. They likely have some redundancy built into the systems, but they're all located in a single location. It's not a single server, but they also didn't appear to exist in multiple geographic locations.
If that's right, to help visualize what's going on, their network might look something like this:
Users and email traffic flow into their data center from the top. What I suspect is broken is some flavor of highly redundant storage that many/all of the email and web servers depend on.
I marked the storage with a red X.
In this scenario they could lose lots of things such as a web server, a mail server, a firewall, a load balancer, probably with minimum impact to the user experience.
The storage I'm describing has a lot of redundancy within it. If a hard drive within it failed that shouldn't cause a problem. A catastrophic failure would have to happen to it. Multiple shelves impacted, several disk groups are no longer available.
Last edited:
@rradam got in with a good reply while I was drawing pictures 
That makes sense if an old storage failed and they were working in migrating to new storage.
Picture a second storage next to the old one, and that unfortunately, some of the user data was only on the old storage (and hopefully backed up!) before the old storage went offline.
That makes sense if an old storage failed and they were working in migrating to new storage.
Picture a second storage next to the old one, and that unfortunately, some of the user data was only on the old storage (and hopefully backed up!) before the old storage went offline.
and this problem has been going on for years so why hadn't they fixed it before it got to a critical stage. Be nice to get rid of the spammers as most of my emails seem to be spam these days most of it is ludicrous but I am intrigued as to who writes this stuff as they cant spell and have no idea of English grammar
Many thanks for that so that means I hopefully wont get all the spam I have been getting when they delete the misused email accounts - and I can buy prescription drugs from a guy in India still who is on safe mail - my B12 injections cost so much where I am and so getting from a reputable seller in India is so much more cost effective but as he runs a legitimate business he won't be affected by any of this
Any guess as to when service will be restored? This has got me in a tough situation. I have no idea as to what kind of time and effort is required to fix something like this. I certainly won't be holding my breath since it took six days to communicate the nature of the problem. Educated guesses are appreciated.
I am afraid the service will not be restored at all in worst case or full data loss happened in best case. One week downtime means total failure.
I'm a bit bored, curious (and tired) myself. taking a guess here, but assuming 10million accounts (most of probably which are free, so limits on mailbox space), 20% loss, guessing no data-dedup in place in their backup/restore strategy, maybe 1PB of data, using LTO6 tapes and two tape drives, ~41 days straight if working non-stop. Many variables to account for though (4 tape drives would almost half the restore time for example, while LTO4 tapes would nearly take twice as long ).
nice to see people thinking clearly about the magnitude of the problem and aren't saying that safe mail has gone for good like so many are. I will put my faith in the knowledge of such as rradam and popowich as I would prefer to remain positive and not dismiss safe mail out of hand.
Agreed and thanks to everyone for their input.
I do think they should give us a rough estimate as to when they expect to be back up. That would certainly help inform me as to how I want to proceed here.
same here tried it earlier this morning and nada all I have is "This website is not available" liked it far better when I actually had a notice up from safe mail saying it was down for critical maintenance. Maybe there is just another problem that has cropped up
