I’ve got an alternative plan of attack. Tonight I am going to take the web server down, bring it up under an old backed up images but disable mysql so old database will not be served or altered and new data lost. Copy the mysql files for nextcloud out, then shut it down, bright up the new, and copy the mysql files for nextcloud in. It is unfortunately slightly more interrupting than using the command to copy the files out of the old image would have been but I can not remember or find it.
I’ve added some additional anti-spam measures, an animated bat cursor, birthdays, a live chat feature, bbcodes for embedding video and a host of other new features to forums.
Please take a look: https://www.eskimo.com/forums
Or under our main website WebApps -> Forums.
Debian failed to mount NFS partitions. Manually mounted them.
MxLinux failed to boot properly, needed to be rebooted again.
All others came up properly.
The last server has been re-re-booted with the correct kernel.
Actually I have to reboot one server again because it has the wrong kernel but nobody seems to be using it yet so should not be customer affecting.
Reboots are completed. I am now checking all servers to make sure they properly mounted NFS partitions and bound NIS properly.
Fail2ban is software installed on all of our servers which detects unauthorized activity and bans the offending IP address. The most common unauthorized activity is failed passwords.
If the same IP is banned twice in a two week period, the IP will be banned for a year. This has been made necessary by huge botnets doing brute force password attacks.
If you get your password wrong, please do not keep trying passwords, give us a call for a password reset.
Fell asleep in my chair, too late to do reboots this morning so will be Sunday morning (midnight – 6am) instead.
After I reloaded Mint recently, I neglected to install rwhod, rusersd, which are responsible for servicing the above commands, and I neglected to allow the necessary ports to connect through the firewall.
All of this has been corrected.
Tonight when the date changes to August 18th Pacific Time, we will be doing another round of reboots. Depending upon whether or not they’ve fixed a problem with the last kernel that caused automatic NFS mounts to fail, this may be a somewhat lengthier than usual process as I’m going to have to check the NFS mounts on each virtual machine after the reboot of each physical host.
If this finishes in a reasonable period of time, virtual private servers will experience a brief outage of from 10-15 minutes each to image the machines (a form of backup).