[MikeDVB]

We're going to be restarting all shared, reseller, and semi-dedicated servers between the hours of 8 PM and 9 PM EST (GMT-5) this evening to load updated kernel and security modules. While we've scheduled the reboot window for 1 hour we expect the individual downtime for each server to be no longer than 5 minutes.

If you have any questions at all, let us know.

[Big Dan]

Thanks for letting us know. You're the first shared host that's ever notified me of reboots in advance. How awesome is that!

[MikeDVB]

One of the kernel upgrades we were planning wasn't quite ready by the maintenance window so we've pushed the reboots back to later this evening. I'm looking at around 12 AM to 1 AM GMT-5 or so. The reboot time will only take 2 to 5 minutes per server.

[JFSG]

Mike, why not use Ksplice? You won't need to reboot this way.

[Blind Bandit]

JFSG, on 29 December 2010 - 04:23 AM, said:

Mike, why not use Ksplice? You won't need to reboot this way.

If I remember correctly mike has been using Ksplice. But not all updates can be done this way. I can only guess the updates where not simply patches and as such needed to be rebooted.

[MikeDVB]

We do use ksplice but we were going to give CloudLinux another try as we've been hearing from other providers that it's much improved and really helps to keep things stable. We can't go from CentOS->CloudLinux without a reboot unfortunately.

We've not made the conversion yet as some quirks that came up on our development server yesterday are still being worked out before we make the changes on our production servers.

[Blind Bandit]

MikeDVB, on 29 December 2010 - 12:37 PM, said:

We do use ksplice but we were going to give CloudLinux another try as we've been hearing from other providers that it's much improved and really helps to keep things stable. We can't go from CentOS->CloudLinux without a reboot unfortunately.

We've not made the conversion yet as some quirks that came up on our development server yesterday are still being worked out before we make the changes on our production servers.

How will this effect the install of Litespeed? As there seems to have been a lot of issues with cloud Linux. Even thought Litespeed supports it.

[MikeDVB]

Blind Bandit, on 29 December 2010 - 12:42 PM, said:

How will this effect the install of Litespeed? As there seems to have been a lot of issues with cloud Linux. Even thought Litespeed supports it.

That is exactly what is supposed to have improved. Several other providers I personally network with have been using CloudLinux on LSWS 4.0.18 and 4.1RC4 with excellent results on servers that are even busier/more densely packed than our own.

We're not going to make the change if it doesn't work 100% and ultimately if it were to cause issues, it's a quick 3~5 minute reboot away from being disabled