VMware Cloud Community
jftwp
Enthusiast
Enthusiast

Curious... who out there reboots ESX hosts monthly or semi-monthly?

Just posing the question in hopes of stirring up a little friendly debate about whether this might not be a bad idea to 'periodically' perform what I'll call 'Maintenance Reboots' of ESX (3.x with Service Console; not embedded 3i), especially those with DRS and HA enabled in clusters.

Thanks.

0 Kudos
18 Replies
virtualdud3
Expert
Expert

I'm not sure why there would be a benefit to performing a scheduled reboot of an ESX server.

My "reboot schedule" is to reboot ESX servers when a patch requires doing so.



###############

If you found this information useful, please consider awarding points for "Correct" or "Helpful". Thanks!!!

############### Under no circumstances are you to award me any points. Thanks!!!
vmroyale
Immortal
Immortal

The patches have pretty much required monthly or every other month reboots for me. I let the hosts run until they need patching.

Brian Atkinson | vExpert | VMTN Moderator | Author of "VCP5-DCV VMware Certified Professional-Data Center Virtualization on vSphere 5.5 Study Guide: VCP-550" | @vmroyale | http://vmroyale.com
jftwp
Enthusiast
Enthusiast

Thanks --- this is already getting good/quick feedback from other 'real admins' out there. Let's keep the comments coming... I'm giving 'Helpful' for everyone until this slows down. Rock on...

0 Kudos
dsanders
Expert
Expert

I don't have a SAN yet so patching my hosts doesn't happen near as often as it should. My hosts have as high as 148 days of uptime now. We were approaching a year of uptime, but we had a power outage 149 days ago Smiley Happy

I also see no reason to setup a reboot schedule.

0 Kudos
rDale
Enthusiast
Enthusiast

Only by force due to monthly patching.

Patching is automated mid month so the servers are patched mid month from last months patches and then rebooted.

0 Kudos
petedr
Virtuoso
Virtuoso

In my experience when I was running a Vmware infrastructure it was much the same. We only rebooted our hosts when needed, not on a schedule. Most or the time it was for patching but one or 2 times a year it would be for other conditions in our data center which required our servers to come down ( power problems, server moves etc ).

www.thevirtualheadline.com www.liquidwarelabs.com
0 Kudos
derekn
Enthusiast
Enthusiast

I only reboot when patching also. Not sure what the benefit of rebooting on schedule does. I have considered restart services in ESX on a scheduled basis.

-go easy
0 Kudos
oreeh
Immortal
Immortal

No scheduled reboots, I only reboot them when needed (due to patching).

0 Kudos
bggb29
Expert
Expert

We also only reboot when patching

0 Kudos
rriva
Expert
Expert

No schedule reboot.

Install patch only if there's a problem and reboot.

Bye

R

If you found this information useful, please consider awarding points for "Correct" or "Helpful". Thanx!

RRiva | http://about.me/riccardoriva | http://www.riccardoriva.com
0 Kudos
jftwp
Enthusiast
Enthusiast

Okay, sounds like everyone is on the same page of "Reboot only when required; usually due to patching". Now the question/discussion morphs into: How often do you patch? Do you patch every time a patch comes out, whether general/security/critical level patch? We usually only patch 'critical' releases, but what do you all consider 'best practices' in this regard? Everyone seems to have a preference/credo where system updates are concerned, so what say 'you'?

0 Kudos
rDale
Enthusiast
Enthusiast

2nd week of every month to pre-production test and then 3 weeks later to production (if all tests are passed)

0 Kudos
bfent
Enthusiast
Enthusiast

We patch to fix a problem or when a patch is released relating to drivers, services, etc that we are using (why fix something that isn't broke). When we patch, we apply ALL patches. This is due to some patches building on previous ones (also VMware best practice). We patch the DR site first then move to production after about a week.

0 Kudos
mikepodoherty
Expert
Expert

To apply patches or when troubleshooting a problem within the Data Center that requires a reboot - such as to check what the HBAs are seeing in terms of the SAN.

I've seen Linux boxes up for over a year between reboots so this uptime isn't surprizing. If these were Windows, then bi-weekly "Therapeutic" reboots would be the SOP.

HTH

Mike

0 Kudos
Curdasss
Hot Shot
Hot Shot

I try and not reboot my ESX servers. I've had one server up for 204 days right now, not a record by anymeans.I just don't reboot them unless they need them, which is not very often.

0 Kudos
khughes
Virtuoso
Virtuoso

We try to only restart our esx servers when we are doing patches/updates. it is definately nice not having to worry about what impact you taking down a host when you can just vmotion everything off to do work on it. unfortunatly you have to reboot the VM servers when you do windows patching =/

-- Kyle "RParker wrote: I guess I was wrong, everything CAN be virtualized "
0 Kudos
jftwp
Enthusiast
Enthusiast

Yeah, but the 'new' ESX/VC 3.5/2.5 patch management at least tries to get its arms around both ESX patching and Windows patching. What it doesn't do, however, is manage VMware Tools upgrades, which is somewhat surprising. They said it will do that in the next release of the patch management feature (of 3.5/2.5).

0 Kudos
khughes
Virtuoso
Virtuoso

we're still running 3.0.2 and 2.0.2. The new features in 3.5/2.5 sound very promising though.

-- Kyle "RParker wrote: I guess I was wrong, everything CAN be virtualized "
0 Kudos