|
I've been tooling around and managing to be the backup to a ~400 VM / 20 host environment for about 8 months now. I somehow got this far without ever getting into the nitty gritty of understanding ESX as an OS itself. Other than installing it and setting up management network I've been pretty content to just manage it via GUI. Troubleshooting an issue where hostd is constantly crashing and not allowing you to connect via management software kind of forced my hand. Its escalated to engineering with vmware support at this point, so it seems like it was indeed over my head after all and I don't need to feel so bad.
|
#
?
May 31, 2013 05:12
|
|
|
|
| # ? May 5, 2024 12:35 |
|
|
warning posted:Troubleshooting an issue where hostd is constantly crashing and not allowing you to connect via management software kind of forced my hand. Its escalated to engineering with vmware support at this point, so it seems like it was indeed over my head after all and I don't need to feel so bad.
|
|
#
?
May 31, 2013 05:38
|
|
|
We don't have the license that has those features. I guess that is enterprise plus? So rebuilding hosts is still a few hours of labor. This isn't the first time we have seen this issue so I have a feeling it is a product of our environment causing it. We rebuilt the last host (twice) to fix it. I had a feeling when the last host needed to be rebuilt the second time it was a persistent issue. Seeing it on another host just confirmed suspicions.
|
|
#
?
May 31, 2013 06:01
|
|
|
Misogynist posted:SRM is a complete waste of money if you already have your storage/data networks stretched to your DR site I'm going to imagine a company that ditched vSphere for XenServer doesn't have stretched networks and storage.
|
|
#
?
May 31, 2013 13:52
|
|
|
Whats the easiest way to clone a vm to a different remote cluster? I was going to use converter because I wanted to sysprep the machine before it booted and apply some different options but twice now Ive run the job and had it error out, maybe because the site I am running the conversion to only has a 10 meg pipe and I had to rate limit? I dunno. Anyways, is that the smartest way? Is there some other obvious method I am missing?
|
|
#
?
May 31, 2013 14:47
|
|
|
Syano posted:Whats the easiest way to clone a vm to a different remote cluster? I was going to use converter because I wanted to sysprep the machine before it booted and apply some different options but twice now Ive run the job and had it error out, maybe because the site I am running the conversion to only has a 10 meg pipe and I had to rate limit? I dunno. Anyways, is that the smartest way? Is there some other obvious method I am missing? Copy the VM directory to the remote datastore then register it as being copied.
|
|
#
?
May 31, 2013 14:55
|
|
|
three posted:I'm going to imagine a company that ditched vSphere for XenServer doesn't have stretched networks and storage. Oh trust me I have made myself heard on that retarded decision and am budgeting for vSphere for our server workloads for 2014. After bringing enough hate to XenServer, I was told that the decision was made because no one here knew vSphere well enough and our network guy had some experience with XenServer. Regaurdless the XenServer pools are not even configured properly. I setup SAN based replication across our two sites, so bringing a datastore online in a site failure is pretty quick. We currently only have a 50mb link between the two sites that is shared with all other traffic. Ongoing replication is pretty quick, but seeding a new datastore with a bunch of data can take a few days. We are getting two pairs of dark fiber handed off to us between our sites in about a month. We are getting the fiber for free actually (thank you state highway department sharing with local government). Initially our plan was just dual 10gb links between the sites. We were also offered a huge discount on carrier grade optical gear, so we are now wading through options for CWDM and DWDM. Shortly I will have intersite bandwidth up the wazoo. Edit: Ha. One of those stupid XenServer hosts hard locked up this morning and took down all the VMs on it. Looks like I may be rebuilding them sooner than later. Moey fucked around with this message at 19:21 on May 31, 2013 |
|
#
?
May 31, 2013 15:21
|
|
|
I just got approval for vSphere Essentials from my boss, for the loving first time ever someone listened when I said that running 40 physical machines that were in our test environment would be the perfect use case for VMware. :woo: A question: These 40 physical machines are broken up into 4 computer nodes, which all need to talk on separate private networks within themselves. Without touching on DV Switches and the like, what is the best way to manage separate private networks across multiple VMware hosts? As far as I know the easiest (and really only way) to do it is to create a new vSwitch that doesn't have any network adapters attached to it, and label it as you want. Is there any better option?
|
|
#
?
Jun 1, 2013 02:05
|
|
|
Wicaeed posted:These 40 physical machines are broken up into 4 computer nodes, which all need to talk on separate private networks within themselves. Without touching on DV Switches and the like, what is the best way to manage separate private networks across multiple VMware hosts? As far as I know the easiest (and really only way) to do it is to create a new vSwitch that doesn't have any network adapters attached to it, and label it as Set servers that need to talk to each other(same network) on the same hosts, what you need to do is make Virtual Standard Switches bound to nics that. Using VMXnet 3 you will have 10gb/s second(assuming the CPU can accommodate).
|
|
#
?
Jun 1, 2013 02:13
|
|
|
or just use a vlan.
|
|
#
?
Jun 1, 2013 02:27
|
|
|
Corvettefisher posted:You are correct, Virtual Standard Switches attached to NICs that are tagged or uplinks tagged to different networks. haha dude, I wish I could use VMXnet3, however these servers run a very custom linux kernal. We are lucky that they had Intel E1000 networking support.
|
|
#
?
Jun 1, 2013 05:00
|
|
|
Is there a clustered Hyper-V backup solution out there that can just backup the OS drive of VMs in the cluster? Preferably one where we can set the rules up to backup all VMs except ones with a particular naming scheme (or some other way to exclude like a custom field)? We provide hypervisor support to various departments on a charge-back model and are evaluating Hyper-V as an option. We currently have specific guidelines for our DR backups (OS drive only and only if less than a certain size). Any backups beyond that (data drives, etc) are something the VM administrators need to configure within their OS. I'm trying to figure out if this is even doable with a Hyper-V backend instead of vSphere like we normally use.
|
|
#
?
Jun 1, 2013 05:42
|
|
|
Fancy_Lad posted:Is there a clustered Hyper-V backup solution out there that can just backup the OS drive of VMs in the cluster? Preferably one where we can set the rules up to backup all VMs except ones with a particular naming scheme (or some other way to exclude like a custom field)? Pretty sure Microsoft's DPM 2010/2012 will do this, the licensing even works in your favor as you only need a System Center license for the hyper-visor and it will cover all the guests unless you end up needing to do point in time restores for exchange/sql.
|
|
#
?
Jun 1, 2013 07:17
|
|
|
NVM figured it out.
Demonachizer fucked around with this message at 14:49 on Jun 1, 2013 |
|
#
?
Jun 1, 2013 14:44
|
|
|
Wicaeed posted:haha dude, I wish I could use VMXnet3, however these servers run a very custom linux kernal. We are lucky that they had Intel E1000 networking support.
|
|
#
?
Jun 1, 2013 15:27
|
|
|
Wicaeed posted:haha dude, I wish I could use VMXnet3, however these servers run a very custom linux kernal. We are lucky that they had Intel E1000 networking support. Wait, what? Also if VMware tools can be installed(which is should) the VMXNet Drivers are installed  says that custom linux kernel is a minimal installation of CentOS or Debian distro. I got touted that a poo poo load with some vApp's then it I typed in init 5, and presto it was just a CentOS install where someone disabled GUI boot.
Dilbert As FUCK fucked around with this message at 21:00 on Jun 1, 2013 |
|
#
?
Jun 1, 2013 20:57
|
|
|
anyone know how to solve the issue of getting prompted to reboot every time you fire up a virtual desktop with paravirtual drivers? I am not sure if it is the disk controller driver or the network, but as soon as we switch to SAS and e1k, the users stop getting prompted to reboot immediately upon login.
|
|
#
?
Jun 1, 2013 21:15
|
|
|
adorai posted:anyone know how to solve the issue of getting prompted to reboot every time you fire up a virtual desktop with paravirtual drivers? I am not sure if it is the disk controller driver or the network, but as soon as we switch to SAS and e1k, the users stop getting prompted to reboot immediately upon login. Sounds like the master image never had it properly installed. Might just want to make sure it isn't doing a prompt to reboot on the master. Also any reason you are using the Paravirtual on VD's? Dilbert As FUCK fucked around with this message at 21:26 on Jun 1, 2013 |
|
#
?
Jun 1, 2013 21:16
|
|
|
Corvettefisher posted:Also any reason you are using the Paravirtual on VD's? While I did not build the image myself, I am quite certain it was fully installed before hand.
|
|
#
?
Jun 1, 2013 23:08
|
|
|
If you still have to install them, they weren't there before.
|
|
#
?
Jun 2, 2013 10:44
|
|
|
evil_bunnY posted:If you still have to install them, they weren't there before.
|
|
#
?
Jun 2, 2013 14:32
|
|
|
adorai posted:They don't have to be installed, they just prompt for a reboot. No matter how many times we reboot the parent image, the clones still prompt. I was assuming it was a hardware identifier issue of some kind, whereby it sees the NIC of the new clone as different hardware or something. Are you sure that's what's causing the reboot prompt? Changing EVC mode prompts for a reboot of the VM (at least it did in 4.1, I haven't changed it since upgrading to 5 and 5.1). Maybe you developed the master in a cluster with a different EVC mode than where you're deploying to?
|
|
#
?
Jun 2, 2013 18:08
|
|
|
Do you have this hotfix installed? The Win7 VMs will detect VMXNET devices as new hardware if you do not. This is all over vmware documentation so most people have it, from what I've seen. http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=1020078
|
|
#
?
Jun 2, 2013 18:25
|
|
|
Erwin posted:Are you sure that's what's causing the reboot prompt? Changing EVC mode prompts for a reboot of the VM (at least it did in 4.1, I haven't changed it since upgrading to 5 and 5.1). Maybe you developed the master in a cluster with a different EVC mode than where you're deploying to? This would be something to look into, if the master image was say, built on amd proc's and you constantly boot onto intel, this error will come up often, or if you have a varied environment of CPu's from different families.
|
|
#
?
Jun 2, 2013 20:54
|
|
|
Erwin posted:Are you sure that's what's causing the reboot prompt? Changing EVC mode prompts for a reboot of the VM (at least it did in 4.1, I haven't changed it since upgrading to 5 and 5.1). Maybe you developed the master in a cluster with a different EVC mode than where you're deploying to? warning posted:Do you have this hotfix installed?
|
|
#
?
Jun 2, 2013 22:06
|
|
|
Crap Nimble is having a VDI Storage workshop the same time Scott low is having a Storage Design Consideration workshop at the Charolette VMUG... Leaning more towards scott lowe
|
|
#
?
Jun 3, 2013 15:47
|
|
|
Corvettefisher posted:Crap Nimble is having a VDI Storage workshop the same time Scott low is having a Storage Design Consideration workshop at the Charolette VMUG... I would go with Scott Lowe. I sat through 8 hours of Nimble training did some joke of a test after. You can learn about CASL on your own pretty easily. And setting up/managing a Nimble is like eating a happy meal.
|
|
#
?
Jun 4, 2013 04:48
|
|
|
Yeah I probably will some of my friends will be in Scotts stuff, Nimbles Storage has a hands on lab though...
|
|
#
?
Jun 4, 2013 04:53
|
|
|
Corvettefisher posted:Yeah I probably will some of my friends will be in Scotts stuff, Nimbles Storage has a hands on lab though... The Nimble labs will literally bore you out of your mind. It is the most simple thing to setup.
|
|
#
?
Jun 4, 2013 04:57
|
|
|
I'm configuring stuff on our new esxi machines now and I'm trying to ensure that a particular partition in one of the virtual machines is aligned with the blocks on the underlying storage. The vmdk will be on local storage formatted with VMFS 5.58 block size 1MB. The storage is a fusion io flash card which is formatted with 512B blocks. I was wanting to create an xfs partition which will be the entire size of the vmdk file. Do I need to bother setting an offset for block alignement or should it just work?
|
|
#
?
Jun 5, 2013 16:47
|
|
|
jre posted:I'm configuring stuff on our new esxi machines now and I'm trying to ensure that a particular partition in one of the virtual machines is aligned with the blocks on the underlying storage. I want to find the document to confirm but this isn't a problem with most modern OS's on VMFS5. It should be 2008 and above and RHEL 6.X on VMFS this is not needed.
|
|
#
?
Jun 5, 2013 17:29
|
|
|
Corvettefisher posted:I want to find the document to confirm but this isn't a problem with most modern OS's on VMFS5. It should be 2008 and above and RHEL 6.X on VMFS this is not needed.
|
|
#
?
Jun 5, 2013 18:07
|
|
|
Its Centos 6.x and I was going to use xfs. I read that if you use mkfs.xfs directly on the disk instead of creating a partition it should be aligned. e.g. mkfs.xfs /dev/sdb edit: added link to source of that info http://www.fhgfs.com/wiki/wikka.php?wakka=PartitionAlignment
|
|
#
?
Jun 5, 2013 18:13
|
|
|
jre posted:Its Centos 6.x and I was going to use xfs.
|
|
#
?
Jun 5, 2013 18:21
|
|
|
jre posted:Its Centos 6.x and I was going to use xfs. Stop. No. Don't do this. Even if you're making a RAID filesystem on top of SSDs for HPC (which that document is speaking about), there are better ways to do this with proper RAID controllers or mdraid flags. Don't use fdisk. DON'T USE FDISK. You should be using parted. Parted is much smarter about these things. CentOS6/RHEL6 uses parted instead of fdisk in Anaconda. It also automatically aligns partitions, and you don't need to worry about it. Don't worry about it. If you're going to worry about it, you need to know a lot more about Linux and partition alignment in general, especially how alignment offsets between VMFS and whatever filesystem you use can cripple performance. Just rely on "parted -a optimal". If you want to be really particular about it, this is a much better guide. You don't need to be really particular about it in ESX. "parted -a optimal" is great. Use "parted -a optimal"
|
|
#
?
Jun 5, 2013 18:24
|
|
|
I'm not using raid, the fusion io storage is a PCI card so there's no disk controllers involved at all.
|
|
#
?
Jun 5, 2013 18:31
|
|
|
jre posted:I'm not using raid, the fusion io storage is a PCI card so there's no disk controllers involved at all. Exactly why you should completely ignore the link you posted and just use "parted -a optimal".
|
|
#
?
Jun 5, 2013 19:13
|
|
|
The filesystem itself isn't responsible for alignment; just ensure you're aligning to a 1MB boundary. So try a partition offset of 128 bytes, or 1MB, instead of the default of 63. Recent editions of fdisk I think also do this, but you should probably start getting used to parted or gparted. See how it does it by modern standards, then mimic it all you want in other tools like fdisk.
|
|
#
?
Jun 7, 2013 13:50
|
|
|
Kachunkachunk posted:The filesystem itself isn't responsible for alignment; just ensure you're aligning to a 1MB boundary. So try a partition offset of 128 bytes, or 1MB, instead of the default of 63. Or don't mimic it in fdisk at all, because you end up using cgdisk or gdisk given that fdisk can't deal with GPT. Like net-tools (ifconfig and friends), fdisk is a thing of the past. Use parted.
|
|
#
?
Jun 7, 2013 15:42
|
|
|
|
| # ? May 5, 2024 12:35 |
|
|
evol262 posted:Exactly why you should completely ignore the link you posted and just use "parted -a optimal". Following your suggestion I've tried the following code:http://lists.opensuse.org/opensuse-bugs/2010-05/msg03671.html code:
|
|
#
?
Jun 7, 2013 16:28
|
|