Some of you, very few of you in fact, may have noticed short-lived disruptions to Kolab Now services over the course of this weekend. This impacts < 1% of our users, really.
Symptoms may include your client to have been disconnected, and maybe asking you to confirm your password. This is inconvenient, but it has had to happen;
- Some of our IMAP back-end servers are allocated limited disk space. When these IMAP back-end servers start running out of disk space, my phone vibrates almost continuously. I don’t like my phone vibrating; It seems like a waste of precious energy.
- Some of our IMAP back-end servers are “old”; They still run Red Hat Enterprise Linux 6. This is a thorn in our eyes, while we aim to move forward and gaze upon the horizon — looking back is just in our way, really.
- A previous iteration of our architecture performed sub-optimally. We had the message spools and mailbox meta-data reside on the same quality of disks, and that’s just a waste of disk space on more expensive resources.
- One of the IMAP back-end servers actually runs CentOS — despite our being done with the “amateur hour” a long time ago — the rest of our systems run Red Hat Enterprise Linux, under support contracts. It really, really needs to go away and die, and more importantly, stop annoying me.
This does include a migration of several terabytes of data. It’ll likely not complete over the course of this weekend (the slowest time of the week for all our customers combined), and so I ambition to improve our process as to be able to tell individual customers on Tuesdays, what might be happening to their mailboxes during the weekend after such Tuesday — so they can get back to us in time and raise objections, should they have justification.
At your service as ever,
Jeroen van Meeuwen