VMware Cloud Community
sgorham
Contributor
Contributor

3.5U3 - any guinea pigs yet?

After the issues that happened with U2 I am curious to see if folks have played with U3 yet. I, for one, won't even think about it for some time

Reply
0 Kudos
53 Replies
khughes
Virtuoso
Virtuoso

I'm downloading it and will play with it in our R/D area but as for production it will be some time, if ever we migrate up to U3, not many things fix/new in this version that benifit us

  • Kyle

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

I think the previous post put the hammer on the nail. TEST IT IN THE LAB/R&D!!!!! I have little sympathy for folks that take bleeding edge releases right into production with no testing. Everyone was quick to blame VMware for not finding a bug when almost none of them bothered to test anything themselves. I for one will test this in my test lab. We have a checklist of items to check when upgrading any OS whether it be Windows or ESX. I highly recommend everyone do the same before they rush to criticize VMware.

That's my 2 cents and beleive me, that's all it's worth.

Reply
0 Kudos
Dave_Mishchenko
Immortal
Immortal

Your post has been moved to the VI: ESX 3.5 forum.

Dave Mishchenko

VMware Communities User Moderator

Reply
0 Kudos
Jasemccarty
Immortal
Immortal

ESX & ESXi 3.5 Update 3, rendered my Dell Optiplex GX620's local datastore useless.

The upgrade (and fresh install) will not support a VMFS file system on my ICH-7 based controller.

To be honest with you, it kind of preturbs me, given that it is a decent box to test on.

Now I've got to go get another controller, or setup another box with an NFS or ISCSI target? That's just silly, especially when 3.5.0, 3.5.0 U1 and 3.5.0 U2 worked fine.

Jase McCarty

http://www.jasemccarty.com

Co-Author of VMware ESX Essentials in the Virtual Data Center

(ISBN:1420070274) from Auerbach

Jase McCarty - @jasemccarty
Reply
0 Kudos
lamw
Community Manager
Community Manager

Upgraded two BL460c with emulex FC's and the systems seems to be running okay, further testing and time.... will tell. Also have ESXi U3 on another Bl460c, haven't loaded any VMs on the system yet. I've heard that vmfs-undelete may have been successful for one individual on the forums, I haven't tried that yet but will be testing that functionality soon enough.

Reply
0 Kudos
donikatz
Enthusiast
Enthusiast

I have little sympathy for folks that take bleeding edge releases right into production with no testing. Everyone was quick to blame VMware for not finding a bug when almost none of them bothered to test anything themselves.

I don't think that's a fair comment re the U2 bug, because it was accidental kill date code that had nothing to do with end-user testing. Plenty of people had it in test for a while, looked stable, rolled live, later the date kill went into effect. It was quite correct to blame VMW for that one.

That said, of course it is important to always put new released into test first. But unless you have a lot of money to truly replicate a prod environment in test, it's usually even better to wait for some serious prod feedback from others as well. Usually no reason to jump headfirst into anything new unless there's something critical.

Reply
0 Kudos
jketron
Enthusiast
Enthusiast

I have upgrades several server so far and not had an issue YET. I did notice that build 120512 in the configuration tab had a Health Status button that showed nice info from the SIM agent on a HP server and in U3 123630 it is gone.

Does anyone know where the Health Status button or info has gone in U3?

Also, I have not had any Vmotion issues between U2 and U3, I did not do fresh installs, I patched all the servers using the patching process and they work fine SO FAR.

Reply
0 Kudos
RParker
Immortal
Immortal

I think the previous post put the hammer on the nail. TEST IT IN THE LAB/R&D!!!!! I

BAH! We upgraded, Production. So what? I always stay ahead of the curve.

I have little sympathy for folks that take bleeding edge releases right into production with no testing. Everyone was quick to blame VMware for not finding a bug when almost none of them bothered to test anything themselves.

And you call yourself a TECH!??! Hahaha.. that's funny... You have little sympathy eh, or maybe you don't have what it takes to know your envirnoment and know how to solve issues, that's what it sounds like to me. What's the matter Col. Sanders, chicken?!?!?

We have a checklist of items to check when upgrading any OS whether it be Windows or ESX. I highly recommend everyone do the same before they rush to criticize VMware.

Good! But don't criticize those that enjoy playing with the NEW stuff when it comes out. If it breaks I can fix it.. I support it, that's what we are here for to LEARN! Someone has to be first, so you over conservative people can know what to expect. I prefer to stay AHEAD of the game rather than hearing it second hand...

Reply
0 Kudos
RParker
Immortal
Immortal

Dell Optiplex GX620

Decent for a Windows box MAYBE, but it's not on the HCL, so all bets are off... That's why we follow the HCL, if it's not there, you can't expect it to be fully supported. . . .

Reply
0 Kudos
jketron
Enthusiast
Enthusiast

I have done some DL380 G5's and DL585 G2

Reply
0 Kudos
Jasemccarty
Immortal
Immortal

Well, given that everything in this world is free... Software, Hardware, etc... I figured I wouldn't be stingy.

And up until I ran the updates, I had 2 SQL boxes on it, 1 biztalk server, 1 virtual desktop, and 2 web servers.

And it ran like a champ.

But again, it is test, and not prod.

Jase McCarty - @jasemccarty
Reply
0 Kudos
Dave_Mishchenko
Immortal
Immortal

They've made some changes in the PCI database that you should be able to work around with a custom simple.map file. The first list is from update 2 and the 2nd is update 3. You'll see that they've started using the IDE driver instead of ata_piix. This is likely to do with the fact that the ICH7 controller is now supported in IDE mode for use with SATA CD-ROMs.

8086 27c0 0000 0000 storage ata_piix 82801GB/GR/GH (ICH7 Family) Serial ATA Storage Controller IDE

8086 27c1 0000 0000 storage ahci 82801GR/GH (ICH7 Family) Serial ATA Storage Controller AHCI

8086 27c3 0000 0000 storage ahci 2801GR/GH (ICH7 Family) Serial ATA Storage Controller RAID

8086 27c4 0000 0000 storage ata_piix 82801GBM/GHM (ICH7 Family) Serial ATA Storage Controller IDE

8086 27c5 0000 0000 storage ahci 82801GBM/GHM (ICH7 Family) Serial ATA Storage Controller AHCI

8086 27c6 0000 0000 storage ahci 82801GHM (ICH7-M DH) Serial ATA Storage Controller RAID

8086 27df 0000 0000 storage ide 82801G (ICH7 Family) IDE Controller

8086:27c0 0000:0000 storage ide

8086:27c1 0000:0000 storage ahci

8086:27c3 0000:0000 storage ahci

8086:27c4 0000:0000 storage ide

8086:27c5 0000:0000 storage ahci

8086:27c6 0000:0000 storage ahci

8086:27df 0000:0000 storage ide

Reply
0 Kudos
Jasemccarty
Immortal
Immortal

I'll give that a try...

Thanks.

Jase McCarty

http://www.jasemccarty.com

Co-Author of VMware ESX Essentials in the Virtual Data Center

(ISBN:1420070274) from Auerbach

Jase McCarty - @jasemccarty
Reply
0 Kudos
Dave_Mishchenko
Immortal
Immortal

Would you be able to update both 27c0 and 27c4 and then post a copy of the file here? I'm sure you won't be the last one to have this problem.

Reply
0 Kudos
reidg1785
Contributor
Contributor

Gave it a shot on one of my ESX servers. VMotioned all the VMs onto another server while updating this one. After it was rebooted and out of maintenance mode I moved some VMs on to it, ones that didn't really matter if they had a problem, and within minutes they just rebooted. Once they are on the server they seem OK. Just out of curiosity, I moved one back to the Update 2 server and it didn't reboot. When I moved it back to the U3 server it rebooted again. Has anybody else experienced this? I couldn't bring both servers to U3 to see if it was just a compatibility issue between U2 and U3 after the VMotion because I didn't want to have the major servers go reboot.

Reply
0 Kudos
Jasemccarty
Immortal
Immortal

Yeah,

We follow the HCL too, in production.

But we also test... Because often times, being on the cutting edge of tech/patches/features... doesn't necessarily agree with current configurations. And that's why you test...

When downtime can cost you or your customers millions of dollars... It doesn't really pay to be on the cutting edge right away.

Cheers.

Jase McCarty - @jasemccarty
Reply
0 Kudos
zik
Enthusiast
Enthusiast

Agreed, but testing would not have uncovered the 3.5U2 bug.

Reply
0 Kudos
Jasemccarty
Immortal
Immortal

I would...

Well, I'm going to try to play with that some today, as the box isn't usable without any VM's, and I can't have any VM's without a datastore...

I'll post what I find.

Jase McCarty

http://www.jasemccarty.com

Co-Author of VMware ESX Essentials in the Virtual Data Center

(ISBN:1420070274) from Auerbach

Jase McCarty - @jasemccarty
Reply
0 Kudos
qmcnetwork
Enthusiast
Enthusiast

Updated 4 ESX servers so far and no issues. I upgraded 3 IBM x346 servers thru VI Update Manager plugin to ESXi 3.5U3. The only thing was I had to run it through remediation 3 times. The first time I think it failed because it tried to apply an older patch (U2 maybe). The second time it applied U3 just fine, but didn't install the tools package. Finally I ran the remediate again to install tools.

On the 4th system, a HP bl460c blade, I remediated to ESX 3.5U3. No issues except it too about 50 to 60 minutes to complete. I thought that was kinda long. I plan on doing the next 6 blades just like it this week.

-Jonathan

Reply
0 Kudos