|
BelDin posted:Not that I know of, but there is a VMFS where you can't have more than 8 hosts connected to a non-NFS datastore that is used for your replica image. That limitation is from View Composer, not VSphere itself. Evidently they hard coded it in composer itself. It was a limitation of vSphere until vSphere 5.1 as well. View 5.1 now also supports up to 32 hosts with VMFS as it was only put in place due to the vSphere limitation. three fucked around with this message at 04:26 on Mar 1, 2013 |
# ? Mar 1, 2013 04:24 |
|
|
# ? May 28, 2024 15:33 |
|
SPICE/QXL, is that worth the troubles of getting it to work, or is it mostly just hot air?
|
# ? Mar 3, 2013 16:52 |
|
three posted:It was a limitation of vSphere until vSphere 5.1 as well. The problem is, you still need to run NFS in order to use the 32 hosts. He sounded like he was trying to avoid NFS. "With View 5.1 and later and vSphere 5.0 and later, an ESXi cluster can contain more than 8 ESXi hosts (up to 32), but you must store the linked-clone replica disks on NFS datastores." Hopefully they fix Composer to support it in the next dot release. VMWare supposedly hard coded the limit in the product because previous VSphere versions didn't support it. View 5.1 planning guide
|
# ? Mar 3, 2013 19:47 |
|
BelDin posted:The problem is, you still need to run NFS in order to use the 32 hosts. He sounded like he was trying to avoid NFS. Looks like it's just a result of View's inability to stay up to date with vSphere improvements (simply shocking that the EUC dept is put on the back burner by VMware). I was wrong, and it isn't supported in 5.1 even though it should now be supported since VMFS can do 32 hosts now: http://www.yellow-bricks.com/2012/11/06/vmfs-file-sharing-limits-increased-to-32/ posted:What the paper doesn’t mention is that in vSphere 5.1 this “file sharing limit” has been increased from 8 to 32 for VMFS Datastores. Cormac Hogan wrote about this a while ago. So to be clear, VMFS is fully capable today of sharing a file with 32 hosts in a cluster. VMware View doesn’t support that yet unfortunately, but for instance VMware vCloud Director 5.1 does support it today.
|
# ? Mar 3, 2013 20:40 |
|
"Horizon" View 5.2 will support clusters of greater than eight hosts, with linked clones on VMFS. http://myvirtualcloud.net/?p=4627
|
# ? Mar 3, 2013 22:23 |
|
Are we expecting that VMware Workstation and/or Fusion will support VT-d at any point in the near future? That is to say, has it been mentioned in any roadmap or quasi-official blog post?
|
# ? Mar 4, 2013 20:58 |
|
Misogynist posted:IBM does this -- without hassle or even a question -- but they don't readily disclose the process. Open up a ticket using the online ESC+ tool. As soon as you have your confirmation that the ticket was submitted and you have a case number, call 1-800-IBM-SERV and get somebody, anybody, on the phone. Once you have a warm body, the second they start to ask you for your personal information, ask to speak with the National Duty Manager. You can escalate your case immediately from there. If you want, they'll often stay on the call with you and the technician until they're satisfied you have the resolution you're looking for.
|
# ? Mar 4, 2013 22:59 |
|
Martytoof posted:Are we expecting that VMware Workstation and/or Fusion will support VT-d at any point in the near future? Not that I am aware of, I believe VirtualBox supports this feature, from what their manual leads. https://www.virtualbox.org/manual/ch09.html#pcipassthrough What are you trying to do?
|
# ? Mar 5, 2013 00:25 |
|
Corvettefisher posted:Not that I am aware of, I believe VirtualBox supports this feature, from what their manual leads. https://www.virtualbox.org/manual/ch09.html#pcipassthrough Nothing quite yet, but the idea of direct i/o sounds promising. This was mainly me trying to decide whether I should buy an i5 3570 which has VT-d but is locked for overclocking ( which I don't do anyway), or the 3570K which has no VT-d but is unlocked for overclocking. I guess if there's no immediate plans to implement VT-d I'll just go with the K variant.
|
# ? Mar 5, 2013 00:36 |
|
There's enough little gotchas with direct IO even when it is supported that I'd say pass on it unless you actually have a use case for it.
|
# ? Mar 5, 2013 00:38 |
|
Martytoof posted:Nothing quite yet, but the idea of direct i/o sounds promising. This was mainly me trying to decide whether I should buy an i5 3570 which has VT-d but is locked for overclocking ( which I don't do anyway), or the 3570K which has no VT-d but is unlocked for overclocking. If you are running workstation, it really isn't worth it; however if you are running ESXI(bare metal) and doing a lab then it might be wise to go with VT-d. Hyperthreading in the 3770 on the other hand might be worth more than VT-d. Dilbert As FUCK fucked around with this message at 00:52 on Mar 5, 2013 |
# ? Mar 5, 2013 00:48 |
|
Allright, thanks for the advice guys
|
# ? Mar 5, 2013 01:48 |
|
All, My company, located in NYC, but with data centers around North America is about to embark on building out a new VMware environment across multiple sites that will host the expansion of our production platform and eventual our dev environments as well. Because this is by far the largest such project I've ever tackled, we'd like to hire a consultant with substantially more VMware experience than I have to provide the kind of guidance and expertise that only years in the field can provide. We already have the hardware (Cisco UCS, four hosts per site, Nimble CS460, Ent+ licensing for production hosts, Ent for Dev) which I'm in the process of building out now. We're looking for someone with the following experience: 1. Multi-site enterprise experience with vSphere and related products and technologies (storage, networking, etc). 2. Ideally dev ops and deployment experience within vSphere. 3. UCS experience would be nice, but not required. I'm not being super specific in my requirements because this is my project and I know VMware and most related technologies well enough to know how useful the person I'm talking to would be, so certifications aren't important. I've had very good luck working with Goons, so I wanted to reach out here and see if there was any interest. I will be doing most of the work-work, and I see no reason why this position can't be handled remotely. The position would be paid on an hourly basis, and because I work absurd hours anyway, it's easy for me to accommodate schedules of people with full time jobs. If you have any interest or questions please PM me or send an email to cesworthy at gmail. Thanks! *(this post was SWSP approved)*
|
# ? Mar 5, 2013 18:06 |
|
Corvettefisher, get paid.
|
# ? Mar 5, 2013 18:10 |
|
Beelzebubba9 posted:2. Ideally dev ops and deployment experience within vSphere.
|
# ? Mar 5, 2013 19:14 |
|
Misogynist posted:Maybe I'm being an obnoxious pedant, but what exactly does devops have to do with vSphere? The DevOps portion is only loosely related (deployment makes more sense), but in the kick off meeting I had two weeks ago, my bosses only had questions about how a new virtual environment could help our development and deployment processes. It's less absurd if you know our work flow, and I know they'll start riding my rear end about this shortly, so I figured I'd try and get ahead of their needs and have answers before they have questions.
|
# ? Mar 5, 2013 19:45 |
|
Misogynist posted:Maybe I'm being an obnoxious pedant, but what exactly does devops have to do with vSphere? My guess would be they want someone who can automate the creation/destruction of the VMs from across all their dc's, and then do the usual devops bits with with puppet/chef/salt. quote:Nimble CS460
|
# ? Mar 5, 2013 23:34 |
|
Martytoof posted:Are we expecting that VMware Workstation and/or Fusion will support VT-d at any point in the near future? Just as an FYI the Windows 8 Client Hyper-V role requires VT-d
|
# ? Mar 5, 2013 23:41 |
|
Nukelear v.2 posted:Must have more years of experience with vmware than our storage company has existed for. Heh. I am currently using two CS-240s and have had no complaints yet.
|
# ? Mar 5, 2013 23:45 |
|
Nebulis01 posted:Just as an FYI the Windows 8 Client Hyper-V role requires VT-d
|
# ? Mar 6, 2013 00:21 |
|
Nukelear v.2 posted:My guess would be they want someone who can automate the creation/destruction of the VMs from across all their dc's, and then do the usual devops bits with with puppet/chef/salt.
|
# ? Mar 6, 2013 00:43 |
|
Misogynist posted:Build automation isn't Devops any more than EC2 is The Cloud. This is driving me nuts. Can you explain for the class what devops is?
|
# ? Mar 6, 2013 00:45 |
|
DevOps is the process that leads to everyone not pulling their hair out when scale of deployment and speed of development releases go up.
|
# ? Mar 6, 2013 00:54 |
|
Combat Pretzel posted:Nested page tables, as in EPT on Intels, not VT-d. Bah, sorry about that, thanks for the correction.
|
# ? Mar 6, 2013 00:54 |
|
Noob question here: if I'm running a computer inside of Virtual Workstation, does that VM expose ANY information about the host machine?
|
# ? Mar 6, 2013 00:57 |
|
three posted:Can you explain for the class what devops is? DevOps is a cultural movement to break down the silo walls between developers and sysadmins (henceforth referred to as Operations). It has strong ties to Agile development practices, though it doesn't necessarily require them. Sometimes the DevOps term is used in place of "agile systems administration," which bothers me a teeny bit less. The idea is that in traditional IT, the componentization of roles means that people pass the buck and don't take responsibility for anything outside of their silo. Developers create a monolithic project and then say to the people responsible for running the thing, "Here's an application bundle and some lovely, half-finished deployment requirements. Make this work by tomorrow." Then, Operations manages to move this thing into production through a massive series of trial-and-error gyrations, so that they can get pages at 3 AM over the next several weeks because no consideration was paid to how the thing is supposed to actually be kept running in production. The debugging instrumentation is usually limited to printf()s and a log file, so at the incident post-mortems, they yell at the developers for writing lovely code and the developers yell at the sysadmins for not knowing how to do their loving jobs. It's also a really lovely way to run a project. DevOps tries to get developers involved in operations, and operations involved in development. In the earliest design stages of a project, operations engineers can say things like "there's no way that we can roll out updates to this infrastructure without downtime," or "it would be really awesome if you could output some metrics whenever you [process a transaction/receive this type of error/flargleglab the flundersnub] so we can really quickly identify where problems are happening at 3 AM." Development says to operations, "hey, all this stuff sounds great, can you make it really easy for us to test our code in something that looks exactly like our production environment so we don't break poo poo on you?" and "what can you do to help us A/B test this feature on a bunch of random users?" The ultimate goal is being able to push new application deployments faster, so that code changes can get into production as soon as they're done to maximize value and you don't waste time just sitting on great new features that nobody can use because they can't make it onto a live server. The endgame is allowing the application's feature set to move at the speed of new business initiatives. If DevOps sounds like a stupid buzzword for "how to run a software project without being a complete retard about it," you're right! There is literally nothing more to DevOps than not running your development-production pipeline like an absolute moron. If you put "DevOps" in a job title, you don't understand DevOps. If you put "DevOps" in a job description or list of responsibilities (and you're not describing your company's culture), you don't understand DevOps. Creating a DevOps team or a DevOps guy or being responsible for DevOps things is making things worse for your organization, not better. It doesn't address any of the problems that DevOps arose trying to address. Bringing this back around: "Automated deployment experience with vSphere" is a really great requirement. Asking someone if they have "DevOps experience" is basically asking someone if they've ever worked in a shop where developers and sysadmins actually talk to each other and don't act like dicks.
|
# ? Mar 6, 2013 01:11 |
|
Above Our Own posted:Noob question here: if I'm running a computer inside of Virtual Workstation, does that VM expose ANY information about the host machine? http://arstechnica.com/security/2012/11/crypto-keys-stolen-from-virtual-machine/
|
# ? Mar 6, 2013 01:16 |
|
Misogynist posted:For the class: A+ information / rant, would read again.
|
# ? Mar 6, 2013 02:32 |
|
Misogynist posted:For the class:
|
# ? Mar 6, 2013 03:25 |
|
Misogynist posted:"Automated deployment experience with vSphere" is a really great requirement. Asking someone if they have "DevOps experience" is basically asking someone if they've ever worked in a shop where developers and sysadmins actually talk to each other and don't act like dicks. What shops are "DevOps" shops anyhow? It would seem almost asinine to attempt to roll out that sort of strategy in a pre-production environment at a company like AT&T, T-Mobile or Verizon. Having used to work for a fortune 500 company as a Sysadmin, it was drat near impossible to get any communication between Devs and our departments until after each box reached the first pre-production checks before go-live. So no, we didn't talk to each other, not unless the box in question didn't pass QC. But that would always result in a flame war. Devs screaming at us over "Improper checking" or "Not being up to date with current best practices." Or members of my department accusing the Devs of "not building with the end-user in mind." The company loves to implement these new information technology strategy initiatives without counting the cost. And DevOps appears to be one of them. I heard that DevOps "training" was being shoved down everyone's throats last year. One of my former co-workers informed me that the "once in a blue moon" flame war was effectively on, at all hours. I believe the drop in morale has led to less productivity, much less the time wasted on continued company training courses. For more info: http://en.wikipedia.org/wiki/DevOps
|
# ? Mar 6, 2013 07:02 |
|
Fuzzyjello posted:What shops are "DevOps" shops anyhow? It would seem almost asinine to attempt to roll out that sort of strategy in a pre-production environment at a company like AT&T, T-Mobile or Verizon. Having used to work for a fortune 500 company as a Sysadmin, it was drat near impossible to get any communication between Devs and our departments until after each box reached the first pre-production checks before go-live. So no, we didn't talk to each other, not unless the box in question didn't pass QC. But that would always result in a flame war. Devs screaming at us over "Improper checking" or "Not being up to date with current best practices." Or members of my department accusing the Devs of "not building with the end-user in mind."
|
# ? Mar 6, 2013 07:18 |
|
Misogynist posted:
Thanks for the post, it was really useful.
|
# ? Mar 6, 2013 19:28 |
|
It seems I've tracked the mysterious crashes of my Win08R2 guests to Trend Micro Deep Security. Their filter driver appears to slowly fill up it's heap space and when that occurs all hell breaks loose. Apparently pre-SP1 it purple screened esx, now it just seems to kill my guests. Their initial response of "have vmware prove it's us" and having now done that, all I get is a way to increase heap size so it takes longer before the issue recurs. Any suggestions for a stable AV solution, kaspersky/mcafee/etc... Staying vshield based would be nice, but this kind of sours me.
|
# ? Mar 7, 2013 16:24 |
|
If you have a VMDK sitting in a >2TB datastore, and you go to expand it - it says the maximum size is 2048GB. Does that account for the -512bytes or not? Can you safely just say expand to 2TB, or do you have to put in 1.99TB or something goofy?
Mierdaan fucked around with this message at 17:32 on Mar 7, 2013 |
# ? Mar 7, 2013 16:51 |
|
Nukelear v.2 posted:It seems I've tracked the mysterious crashes of my Win08R2 guests to Trend Micro Deep Security. Their filter driver appears to slowly fill up it's heap space and when that occurs all hell breaks loose. Apparently pre-SP1 it purple screened esx, now it just seems to kill my guests. Their initial response of "have vmware prove it's us" and having now done that, all I get is a way to increase heap size so it takes longer before the issue recurs. What version of DS are you on? I just upgraded to 8.0SP2 and it seems to have resolved quite a few issues. We're running ESXi5.0U2 here. Though it's still a buggy piece of poo poo...
|
# ? Mar 7, 2013 17:03 |
|
Mierdaan posted:If you have a VMDK sitting in a >2TB datastore, and you go to expand it - it says the maximum size is 2048GB. Does that account for the -512KB or not? Can you safely just say expand to 2TB, or do you have to put in 1.99TB or something goofy? Are you expanding the VMDK or the DS volume? VMFS5 The max VMDK size is 2TB-512 bytes The max Luna size is 64TB
|
# ? Mar 7, 2013 17:20 |
|
I'm expanding the VMDK. I know the max size is 2TB - 512bytes, I just don't know if the UI takes that into account and subtracts 512bytes for you, or if you need to put in 1.99TB or 2047GB or something to account for it.
|
# ? Mar 7, 2013 17:32 |
|
IDGI. I *like* IP based. Why do you hate freedom/america? e: vmware doc is spectacularly comprehensive and unhelpful. The dvportgroup is configured for IP-based. The Nexuses are configured for a normal etherchannel (and it works peachy with a VSS, but not a VDS). evil_bunnY fucked around with this message at 18:26 on Mar 7, 2013 |
# ? Mar 7, 2013 17:54 |
|
Goon Matchmaker posted:What version of DS are you on? I just upgraded to 8.0SP2 and it seems to have resolved quite a few issues. We're running ESXi5.0U2 here. Though it's still a buggy piece of poo poo... 8.0 SP1, none of the filter driver changes for SP2 seem to be related and support didn't think it would help. 9.0 quadruples the default heap size, so that's a bit telling. Issue appears to be related to the network portion of DS, even though we aren't licensed or using it, every connection get entries in that heap. At their suggestion I'm going to change the firewall over to tap mode, might help, I didn't even look at it since we don't use it. Really their solution seems like a band-aid and when it was just a buggy appliance or agent, I could deal with it, but glitchy kernel level code in ESX is pretty untenable to me.
|
# ? Mar 7, 2013 18:27 |
|
|
# ? May 28, 2024 15:33 |
|
In Hyper-V do I want to shut down the VMs when patching the physical server?
|
# ? Mar 8, 2013 20:08 |