FastMail.FM Status

The following changes have been rolled out to production:

• The keyboard shortcut to select all has changed: it is now "* a", which means press the star key (normally shift-8), then after that press the "a" key within 1 second. You can deselect all using "* n" (star followed by the "n" key). The previous shortcut of just "a" on its own conflicted with the shortcut for reply to all when the reading pane is active, and it was also too easy to accidentally select all before hitting another key which could then perform an action on all those messages.

We experienced another hang on one of our servers. As noted in a previous blog post, we’re in the process of making crashdump work so that this behaviour can be diagnosed and fixed.

The crashdump didn’t quite succeed, but the subsequent reboot did, so that the server was back up and running in about five minutes.

All is operating normally again now.

Status (by brong at Sun May 12 10:50 UTC)
One of our servers is heavily loaded and responding more slowly than usual. We are investigating.

Status (by brong at Fri May 10 22:58 UTC)
We have been having intermittent network issues in our New York datacentre for the last half an hour. This will have caused the web interface to appear slow, and maybe delayed email as well.

The servers themselves have been running fine the whole time, it’s just the network which is having issues.

Update (by brong at Sat May 11 00:03 UTC)
Update from NYI: there was a DDOS attack against another of their customers which overloaded the links that we share. They have now mitigated that traffic, so our network is working smoothly again.

Status (by brong at Fri May 10 17:59 UTC)

Update (by brong at Fri May 10 18:42 UTC)
Everything is running again

Status (by robm at Thu May 9 04:33 UTC)
A change rolled out about 2 days ago broke push updates of new emails in the new user interface. This has now been fixed, so new emails should now be pushed to the new interface when you’re logged in again.

Status (by brong at Wed May 8 05:03 UTC)
one of our servers is currently experiencing very high load – moving users off it

Update (by brong at Wed May 8 05:08 UTC)
All users moved to other servers – the cause was high IO patterns, once they are finished we can bring that server back into regular service.

Everything should be running smoothly again now.