This evening I will be booting ice into a new kernel to test NFS.

      I am working with a Linux kernel developer to debug some kernel nfs server issues that are new to the 5.7.x kernels.  I plan to boot into the new kernel around 11pm, then things may or may not be interrupted for some time if NFS does not work while I gather information that may be helpful to the kernel developers before reverting to a known good kernel.

     This will affect mail, and also vps1 through vps7 but not higher numbered virtual  machines.  The virtual machines will only be down for minutes during the boot, mail be unavailable for a longer period as it is what is affected by the NFS problems and it may take some time to gather all of the necessary information to allow a fix to be developed.

     Between 11PM Sunday June 21st and 2AM Monday June 22nd:

     vps2, vps3, vps4, vps5, and vps7 about 1/2 hour each for imaging.

     scientific7, uucp, and mx1 about 1/2 hour for scientific7 and about 15 minutes for the other two.  mx1 will not be service affecting as mx2 will handle the traffic while it is down.  You can use centos7 as an alternative to scientific7 during it’s downtime, they are essentially the same code base.

     mint, ubuntu about an hour each, ftp/web about 1/2 hour.

     Mint and Ubuntu have similar code base to Debian, Julinux, and Zorin, suggest using one of those as alternate during this work.

     Webmail, imap, and pop were unavailable from 7:15PM to approximately 8:30PM.

     I restarted mail to test a new kernel before I restart the large servers tonight.  For some reason dovecot automatic start failed and I did not realize it at the time.

     I got a call just before 8:30, checked, and found it was not running and restarted and it started fine so no idea what went wrong earlier.

     An update; I discovered what went wrong earlier is that at some point Ubuntu changed where dovecot / systemd expected the PID file to be and it required manually editing the unit file.  I did that so it should be okay in future restarts.

     Saturday morning just after midnight Pacific time, I will be rebooting servers into a new kernel.  If this kernel behaves properly we will remain on 5.7.4, if not I will need to reboot again to 5.6.0.  5.7.0-5.7.2 has not functioned properly as NFS server.

     The reboot did not in and of itself fix the problem but everything restarted did point me at the source so I could fix it.  It is completed, everything should be functional again.

     I have to reboot a machine because the NIS subsystem, which is how authentication information is pushed to all machines, is not working properly on this machine.

     This will briefly make e-mail unavailable as well as the shell servers centos7, centos8, scientific7, and fedora.

     I will be rebooting all servers shortly after 1AM to load a new 5.7.2 kernel, the 5.7.0 kernel was slightly unstable in nfs server applications.

     The shell server, mxlinux.eskimo.com, is again available.

     It has been upgraded from 18.3 to 19.2.  So far 19.2 appears to be MUCH more solid and bug free relative to 18.3.  I have yet to find any actual bugs in the implementation (but a hell of a lot of operator errors).

     I have not installed many user applications yet, so many mailers, news readers, editors, etc, are not there yet but will be soon.  I think you may want to consider this as a good alternative to debian and ubuntu now.  It is based off of the Debian 10 code base.

     The mail system is basically functional and alpine and mailx are installed but most other mailers are not yet installed.  The pan news reader is installed, most others are not installed.

     Iglulik spontaneously rebooted about 3AM, this is the second time it has done this and Ice, the server which has /mail spool, has also spontaneously booted once.  It appears that 5.7 has some stability issues when NFS file systems are exported.  The machines not exporting file systems have been completely stable.

     I know that the Linux community is doing a lot of work on NFS right now, which is good because NFS in Linux has been buggy forever but not in ways that crash machines but in ways that when a server goes away and comes back the clients do not always recover properly.  There was a lot of work in the 5.7 kernel and they just came out with a new nfs-kernel package.  So hopefully they’ll get this resolved soon.

     So I’ve compiled 5.7.1 now on Iglulik.  If it does not spontaneously boot again before tonight I will be rebooting in the early AM to load the new kernel.

      I am taking mxlinux down to upgrade from MxLinux 18.3 to MxLinux 19.2.  Because of a change in the debian code base upon which MxLinux is based and because they’ve bastardized Debian beyond the point where normal upgrade procedures can be used, a complete re-install is necessary so it may be down for a few days.