|
BangersInMyKnickers posted:You've probably worked around it already but throw it in 2003 compatibility mode or figure out where the exe extracted the msi to (usually appdata\local or locallow) and try running that directly. I figured out the error but this isn't what the error should say. I remote'd into vdc.vm.local(virtual Domain Controller) and not my vc.vm.local(vCenter Server), took me about 15 minutes before I realized I was being a dummy.
|
#
?
Apr 30, 2013 17:45
|
|
|
|
| # ? May 12, 2024 19:03 |
|
|
Internet Explorer posted:I'm fairly sure he knows that. I would hope so.  How is the View 5.2 testing going CF? Wondering how long I should wait before beginning the upgrade process on our 5.0.1 environment.
|
|
#
?
Apr 30, 2013 18:40
|
|
|
Moey posted:I would hope so. Depends what ESXi/vCenter version are you running, Host cache and CBRC breaks on some versions, also are you using CA certs on hosts/vcenter, or the default certs? So far I really like it, much better than the 5.0 to 5.1 jump. There are just some really quirky bugs you need to look out for. Most bugs I have come across are with vCenter/SSO, no surprise there.
|
|
#
?
Apr 30, 2013 18:49
|
|
|
Anyone know if it's possible to monitor the Network Adapter Type inside a VMware guest? I've got a really annoying issue with a Virtual Appliance. Basically they all share the same hostname so if a user switches it from NAT to Bridged everyone in the company will hit their VM when using a software client from the host. I'm looking into if there's a way to force the DHCP domain but if I could just write a small service that throws up an alert when it detects the VM is set to bridged I think I'll be okay.
|
|
#
?
Apr 30, 2013 18:58
|
|
|
Hooray vSphere 5.1 U1 supports SQL 2012 now! http://www.vmware.com/support/vsphere5/doc/vsphere-vcenter-server-51u1-release-notes.html#whatsnew https://www.vmware.com/support/vsphere5/doc/vsphere-esxi-51u1-release-notes.html
|
|
#
?
Apr 30, 2013 20:38
|
|
|
Corvettefisher posted:Hooray vSphere 5.1 U1 supports SQL 2012 now! It also lets you set the VMFS3.MaxHeapSizeMB value to 640MB. Handy when you've got more than 25TB worth of VMs powered up on a host. Ran into an issue this past weekend where that wasn't enough... http://kb.vmware.com/kb/1004424
|
|
#
?
Apr 30, 2013 21:05
|
|
|
Oh hey I guess I should upgrade to 5.1 Somehow I forgot to do that
|
|
#
?
Apr 30, 2013 21:33
|
|
|
Number19 posted:Oh hey I guess I should upgrade to 5.1 Research up SSO requirements PRIOR to doing, there are some GOTCHA's depending on the size, and complexity of your environment. It is very different from 4.0=>4.1 or 4.1=>5.0
|
|
#
?
Apr 30, 2013 21:39
|
|
|
The biggest gotcha is "this loving thing never, ever works"
|
|
#
?
Apr 30, 2013 21:50
|
|
|
Corvettefisher posted:Research up SSO requirements PRIOR to doing, there are some GOTCHA's depending on the size, and complexity of your environment. It is very different from 4.0=>4.1 or 4.1=>5.0 The environment is small and I got the vCSA up and running when I was testing it so I should be fine.
|
|
#
?
Apr 30, 2013 22:12
|
|
|
Now I remember why I didn't upgrade...it's because Backup Exec doesn't support 5.1 yet. I checked before starting my upgrade and the latest update says they probably won't support it until July. 5.1 was released at the end of August 2012. I really didn't need any more reasons to toss Backup Exec out the door but this pretty much salts it. It's taking them almost a year to support for it. Great job guys 
|
|
#
?
May 1, 2013 00:34
|
|
|
Just a question what do you all use for Backing up? I'm a fan of PHDvirtual myself, I know Veeam is pretty frowned upon which my new place uses but has many gripes about. Also the new Symantec A/V for VDI is actually pretty loving good.
|
|
#
?
May 1, 2013 01:17
|
|
|
Corvettefisher posted:Just a question what do you all use for Backing up? Was only using Veeam at my old place. It worked and I passed a full DR test with flying colors because of it. It did have quirks and would poo poo the bed randomly and want to do a new full backup on random VMs. New place we are just starting to use PHD Virtual for local backups, then SAN based snapshots and replication to a different site. Quick question on an iSCSI NAS and MPIO vs Teaming the NICs. Lets say I have a very simple (and single point of failure environmet) that has 4 hosts each with a 1gb nic for iSCSI. These nics are connected to a storage switch. The NAS has 4gb nics all connected to the same storage switch. Is there a disadvantage to teaming together all 4 of the NAS nics vs 4 single 1gb links along with multipathing?
|
|
#
?
May 1, 2013 04:07
|
|
|
Moey posted:
Before I answer are you using the Host licensing or the View licensing? Dilbert As FUCK fucked around with this message at 04:14 on May 1, 2013 |
|
#
?
May 1, 2013 04:11
|
|
|
Corvettefisher posted:Just a question what do you all use for Backing up?
|
|
#
?
May 1, 2013 04:12
|
|
|
Corvettefisher posted:Just a question what do you all use for Backing up? I use Veeam. Sometimes it sucks and it has some annoyances, but we literally run production out of the DR site for a whole day every quarter with zero issues and easy failback.
|
|
#
?
May 1, 2013 04:15
|
|
|
Corvettefisher posted:Before I answer are you using the Host licensing or the View licensing? Host running standard. My old job just deployed a new nas and teamed together 4 connections on it. I always have done multiplathing but they are currently going through a single switch. Edit: Single switch because they sprung for 10GbE but don't see in spending the money for a second 10GbE switch for the storage network... Moey fucked around with this message at 04:41 on May 1, 2013 |
|
#
?
May 1, 2013 04:32
|
|
|
At my old job we used VMware Data Recovery which was the worst poo poo. I basically treated it as a nice-to-have but totally unreliable bonus if it was actually working at any given time and paired it with slightly less lovely guest OS agent-based backups. I hear (from this thread) that the newer Avamar-based backup tools are far, far better, but my current place is on ESXi 4.0 with lapsed support, so welp.
Docjowles fucked around with this message at 04:50 on May 1, 2013 |
|
#
?
May 1, 2013 04:45
|
|
|
Moey posted:Host running standard. MPIO will ensure that you get an even distribution across all 4 of your links, while teaming might give you a bad split depending on your hashing algo. MPIO is worth it, in my opinion, since it's mostly set/forget. Also, it sets you up for the future when you might get access to another switch or a 10Gb switch, etc.
|
|
#
?
May 1, 2013 07:24
|
|
|
Corvettefisher posted:Hooray vSphere 5.1 U1 supports SQL 2012 now! I'm still waiting on Postgres support for external databases so I can ditch SQL Server entirely.
|
|
#
?
May 1, 2013 12:31
|
|
|
Corvettefisher posted:Depends what ESXi/vCenter version are you running, Host cache and CBRC breaks on some versions, also are you using CA certs on hosts/vcenter, or the default certs? Yea I went through the vCenter 5.1 upgrade at my old place, that was a joy. Looks like I don't currently don't have to upgrade our vCenter to use View 5.2 though.  Right now our hosts and vCenter are all running self signed certs, but I do have a root CA deployed for the domain, so replacing the certs shouldn't be an issue? madsushi posted:MPIO will ensure that you get an even distribution across all 4 of your links, while teaming might give you a bad split depending on your hashing algo. MPIO is worth it, in my opinion, since it's mostly set/forget. Also, it sets you up for the future when you might get access to another switch or a 10Gb switch, etc. I didn't even think about the hashing algorithm. As for the future once they get another 10GbE switch (or finally decide to converge with the 10GbE backbone that has more than enough switch ports), never though of that. Good call! I would prefer it to be setup properly before too much production stuff gets migrated to it. I am still doing consulting for them and normally handling the VMware stuff. It's always fun to get connected and see what they have mucked up from week to week. Moey fucked around with this message at 15:45 on May 1, 2013 |
|
#
?
May 1, 2013 15:26
|
|
|
Number19 posted:Now I remember why I didn't upgrade...it's because Backup Exec doesn't support 5.1 yet. I checked before starting my upgrade and the latest update says they probably won't support it until July. 5.1 was released at the end of August 2012. We're moving to AppAssure for our backups since it just replicates blocks directly from the OS and doesn't have the dependency/integration headaches. Plus some other cool stuff like its ability to stream the data to your DR site so stuff is already staged in their inventory. Nice for us since we aren't big enough to justify two production clusters with cross-replication.
|
|
#
?
May 1, 2013 18:39
|
|
|
Moey posted:Yea I went through the vCenter 5.1 upgrade at my old place, that was a joy. Looks like I don't currently don't have to upgrade our vCenter to use View 5.2 though. vCenter and Hosts can run without replacing the Certs but you'll need to generate valid CA certs for View connection servers, it will use vCenter/Composer/Secuirty Server self sign but it might yell at you. Also if your hosts aren't 5.1 then some of the new vSGA features, which are a lot better than previous, won't be available till you update.
|
|
#
?
May 1, 2013 19:06
|
|
|
Corvettefisher posted:vCenter and Hosts can run without replacing the Certs but you'll need to generate valid CA certs for View connection servers, it will use vCenter/Composer/Secuirty Server self sign but it might yell at you. Also if your hosts aren't 5.1 then some of the new vSGA features, which are a lot better than previous, won't be available till you update. Thanks. Right now I am using a linux based load balancer to point at my connection servers. I plan on bringing up new connection servers anyway as this environment seems to always have strange quarks (mix of misconfiguration that I am hunting down as well as a network built as stable as a house of cards).
|
|
#
?
May 1, 2013 19:11
|
|
|
Anyone have an easy way to calculate how much a VM costs? It doesn't have to be accurate for chargeback purchases, more of a 'look VM's aren't free and can't just be provisioned willy nilly'. Is there an accepted formula? Say my main cluster has about 600K of expenses tied up into it and someone wants 12 average sized VM's.
|
|
#
?
May 1, 2013 19:36
|
|
|
Corvettefisher posted:Just a question what do you all use for Backing up?
|
|
#
?
May 1, 2013 20:40
|
|
|
skipdogg posted:Anyone have an easy way to calculate how much a VM costs? It doesn't have to be accurate for chargeback purchases, more of a 'look VM's aren't free and can't just be provisioned willy nilly'. Is there an accepted formula? Say my main cluster has about 600K of expenses tied up into it and someone wants 12 average sized VM's. Here's one quick way to do it (ignoring that the 600k is a sunk cost :p): $600k / (expected lifespan of that investment in months) / the resource that is important to you (memory?disk?) * the amount they want to use $600k / 60 (say 5 years) = $10,000/month Say you have 1000GB (or whatever) of memory and that is your benchmark for usage -- $10,000 / 1000GB * (12 VMs * 8GB/ea) 10000 / 1000 * 96 = $960/mo (96gb = roughly 10% of your available memory, so on a per month basis it accounts for approximately 10% of the system) Or all on one line - $600,000 / 60 / 1000 * (12 * 8) = $960/mo.
|
|
#
?
May 1, 2013 21:04
|
|
|
BangersInMyKnickers posted:We're moving to AppAssure for our backups since it just replicates blocks directly from the OS and doesn't have the dependency/integration headaches. Plus some other cool stuff like its ability to stream the data to your DR site so stuff is already staged in their inventory. Nice for us since we aren't big enough to justify two production clusters with cross-replication. AppAssure is part of a Dell pitch that someone's making for us for later this year. I'm going to get a demo of it soon but I've heard terrible stories about their support. When are you putting it into place? It's also going to have Force10 switches in it which I don't know a lot about. I've heard some good stuff but not a lot in general. They will be replacing some mid-level Procurve switches which have been ok.
|
|
#
?
May 1, 2013 21:09
|
|
|
BangersInMyKnickers posted:We're moving to AppAssure for our backups since it just replicates blocks directly from the OS and doesn't have the dependency/integration headaches. Plus some other cool stuff like its ability to stream the data to your DR site so stuff is already staged in their inventory. Nice for us since we aren't big enough to justify two production clusters with cross-replication. It still relies on VSS/Shadow copy, if the writers are having a bad day your backups will fail. We have had to deploy VSS hotfixes from Microsoft to get some stuff working.
|
|
#
?
May 1, 2013 21:25
|
|
|
Can anyone say whether or not Xen changes the UUID of a VM when it's migrated? We have a vendor whose lovely ecommerce site apparently is tied to the machined UUID, and they say it changed last night. The hosting service we're using is running Xen, which I haven't ever used.
|
|
#
?
May 1, 2013 21:57
|
|
|
Number19 posted:AppAssure is part of a Dell pitch that someone's making for us for later this year. I'm going to get a demo of it soon but I've heard terrible stories about their support. When are you putting it into place? I can't speak to the switches, but we've been using their PowerConnect stuff for years and it's been good to us. I've run the AppAssure demo for about a month and hit a problem when it stopped talking to clients but I had one of their support guys on the phone and he fixed me up in maybe 30 minutes. It is dependent on VSS, but this also means its really good about knowing exactly what data has been changed and what needs to be replicated, and if it is a redundant block that has already been sent in by a different VM in the repository (think OS patching) it will skip it and just throw up the pointer. You don't see the big bursts in IOPS like you do from a more traditional backup solution and being able to get hourly granularity without a big storage or performance cost is really cool. The dedupe rate has been around 70% but I haven't thrown any of our heavy stuff like the doc management system or Exchange at it yet. For VM's, licensing is per host processor which works out much better than BackupExec for us. We'll probably go live with it in a month or two. Our backup guy just quit so we're replacing him and I'm just doing the proof of concept work before I hand it off. e: One issue at the moment, it doesn't support GPT boot volumes for UEFI systems (which I assume also means the latest 5.1 VM hardware). That's getting addressed with a new release sometime this quarter. BangersInMyKnickers fucked around with this message at 22:32 on May 1, 2013 |
|
#
?
May 1, 2013 22:30
|
|
|
sanchez posted:It still relies on VSS/Shadow copy, if the writers are having a bad day your backups will fail. We have had to deploy VSS hotfixes from Microsoft to get some stuff working. What version of Windows were having the problems? Most of our backups right now are BE agents which runs off VSS and the only time we've had grief was with 2003 servers.
|
|
#
?
May 1, 2013 22:32
|
|
|
BangersInMyKnickers posted:What version of Windows were having the problems? Most of our backups right now are BE agents which runs off VSS and the only time we've had grief was with 2003 servers. 2008 R2 needed a hotfix to work with the version we installed a few months ago. It was a Hyper-V environment which might have made a difference, the domain controller guests needed a patch for the AD backup to work. It's a nice product all in all, we've started trying to sell it over Veeam. I hope dell doesn't screw it up. It does seem to throw errors like mount check and replication failures from time to time which resolve themselves without any intervention. The latest update did help with those. sanchez fucked around with this message at 22:43 on May 1, 2013 |
|
#
?
May 1, 2013 22:40
|
|
|
Mierdaan posted:Can anyone say whether or not Xen changes the UUID of a VM when it's migrated? We have a vendor whose lovely ecommerce site apparently is tied to the machined UUID, and they say it changed last night. The hosting service we're using is running Xen, which I haven't ever used. We have a few XenServer clusters here. Just migrated a VM between hosts and verified the UUID is the same. Don't see why it would be any different with Xen.
|
|
#
?
May 1, 2013 23:08
|
|
|
Mierdaan posted:Can anyone say whether or not Xen changes the UUID of a VM when it's migrated? We have a vendor whose lovely ecommerce site apparently is tied to the machined UUID, and they say it changed last night. The hosting service we're using is running Xen, which I haven't ever used. A couple questions come to mind that might help people answer your question: - How exactly did you migrate it? Some type of live migration, shut down and copy files, etc? - That's weird that the software is tied to a UUID. Is the vendor SURE it's not tied to something else? I mean, if it's tied to the UUID how would the software function in a non-Xen environment or on bare metal?
|
|
#
?
May 1, 2013 23:21
|
|
|
I can only speak for XenServer, and not at the level of a hosting environment, but being tied to the UUID sounds really weird. I'm not even sure you can query the UUID of the VM from within a guest. The UUID is tied to the VM object. So the only way the UUID should change is if the VM object changes. One way would be if the virtual disk was removed and then added to a new VM. Another would be if if it was exported and then imported to be moved to a new pool. I can't really think of any other ways the UUID of a VM would change.
|
|
#
?
May 1, 2013 23:36
|
|
|
Is it just me or did 5.1 Update 1 drastically reduce the ram usage for SSO/inventory/vCenter? I am not even hitting 2 gigs now.
|
|
#
?
May 2, 2013 00:08
|
|
|
Corvettefisher posted:Is it just me or did 5.1 Update 1 drastically reduce the ram usage for SSO/inventory/vCenter? I am not even hitting 2 gigs now. poo poo, guess I better get VUM working then...
|
|
#
?
May 2, 2013 01:31
|
|
|
FISHMANPET posted:poo poo, guess I better get VUM working then... Yeah after a composer install I am still sitting around 2.5Gb, not sure if it matters any but I did also install Java7u17 as well
|
|
#
?
May 2, 2013 02:33
|
|
|
|
| # ? May 12, 2024 19:03 |
|
|
I just took this stupid Prove It online test for a gig, the test was for VMware. Workstation version 7. Apparently an ancient version. Ugh. Somehow managed and 80% (average was 54%) and the questions were two steps from retarded. Not bad considering I have only used ESX server 3.5 and 4.
|
|
#
?
May 2, 2013 02:34
|
|
