The control panel (SolusVM) that manages the nodes has managed to generate over 30,000 processes stuck in wait, all taking up CPU time. this has been reported upstream as a bug.
A fix has been launched however killing 30,000 LVM based processes has put the server under significant further load, it will be left for 30 minutes to complete however if at 21:00 (UK/London) it has not completed a view will be taken on the likely time left and the server will be rebooted if required.
Sorry for the inconvenience.
Update: 28th November 2020 @ 21:23 - it looks like it could take upto 2 days with seriously degraded performance to resolve using the supplied fix as such the server is being rebooted under emergency maintenance.
Update: 28th November 2020 @21:39 - All services have been started, due to the unclean restart many guest servers will be running fsck, this should be fairly brief and service will return to normal.
Saturday, November 28, 2020
