Kernel 4.14 works on one VPS but not the other

[hakkuo23]

So the VPS is running on kernel 4.12, and doing an upgrade of that config to 4.14, the VPS won't boot and shows this:
https://i.gyazo.com/f806f1031f02aa77ec2 ... c51970.png

However, with that same config, it boots just fine on another VPS with the same hypervisor type.

I'm not sure how to debug, and has anyone experienced anything similar to this?

Thanks!

[Edge100x]

That output doesn't show a crash or other problem -- it has kernel output that we'd expect on a working system.

Some OS versions show kernel output when the system is taking a long time to boot or there are other messages that show warnings (and not otherwise). Sometimes such output is overlaid on top of a usable login prompt. Sometimes it is before a different prompt for input about a disk-related error. Are there more lines after this? Have you tried pressing keys (such as enter) and making sure that the system is actually down?

All of our VDSes use the same hypervisor here (Xen). Are you saying that this is on a dedicated machine that you have your own hypervisor installed on?

[hakkuo23]

That output doesn't show a crash or other problem -- it has kernel output that we'd expect on a working system.

Some OS versions show kernel output when the system is taking a long time to boot or there are other messages that show warnings (and not otherwise). Sometimes such output is overlaid on top of a usable login prompt. Sometimes it is before a different prompt for input about a disk-related error. Are there more lines after this? Have you tried pressing keys (such as enter) and making sure that the system is actually down?

All of our VDSes use the same hypervisor here (Xen). Are you saying that this is on a dedicated machine that you have your own hypervisor installed on?

Yeah I spammed the keyboard and let it sit for 2 hours but it didn't budge. Rebooting and choosing 4.12 lets it boot as normal. This is occuring on a VPS.

[Edge100x]

Without the error, it will be hard to troubleshoot this remotely.

You might see if there's an experimental kernel available for your OS and test that, or try updating the OS in other ways. You could check /var/log/messages (or wherever your distribution stores logs). You could check the kernel command line to make sure that it references the correct boot device. You could try not mounting the SSD as part of the boot, if you have one and you are doing that. You could ask us to move your VDS to a different machine (if this is one that is run on our system; that's still not clear, since you said you were running a different hypervisor and didn't answer my question about it).

[hakkuo23]

Without the error, it will be hard to troubleshoot this remotely.

You might see if there's an experimental kernel available for your OS and test that, or try updating the OS in other ways. You could check /var/log/messages (or wherever your distribution stores logs). You could check the kernel command line to make sure that it references the correct boot device. You could try not mounting the SSD as part of the boot, if you have one and you are doing that. You could ask us to move your VDS to a different machine (if this is one that is run on our system; that's still not clear, since you said you were running a different hypervisor and didn't answer my question about it).

Oh all the VPSes are Gentoo, all are on nfo, but all are in different locations. The ones acting up are in Frankfurt and Seattle.

[Edge100x]

It's worth looking at the things I mentioned, and others. Do they all have the exact same old kernel and old kernel boot line?

Gentoo is a good distribution but it requires much more expertise, and work, to install and maintain than Ubuntu, Debian, or CentOS. Troubleshooting strange problems associated with significant upgrades is unfortunately not uncommon with Gentoo.

[hakkuo23]

It's worth looking at the things I mentioned, and others. Do they all have the exact same old kernel and old kernel boot line?

Gentoo is a good distribution but it requires much more expertise, and work, to install and maintain than Ubuntu, Debian, or CentOS. Troubleshooting strange problems associated with significant upgrades is unfortunately not uncommon with Gentoo.

It's odd but I found the issue.

For all the VPSes I set the /etc/default/grub to set the IO scheduler to BFQ:
GRUB_CMDLINE_LINUX_DEFAULT="elevator=bfg numa=off"

And the first one has no issues with this, but I guess commenting out that line in the other two VPSes causes it to boot just fine.

I've just made a script in /etc/local.d to set the IO scheduler from now on.