|
Martytoof posted:Isn't that what I wrote? I think the confusion is about running it on bare metal. You can take Apple hardware, install ESX on it, then run OSX vms on it. You don't have to run the vms on Fusion which has OSX running directly on the hardware.
|
#
?
Oct 8, 2013 19:23
|
|
|
|
| # ? Apr 29, 2024 02:36 |
|
|
I think we're all saying the same thing  The hardware just needs a gray and/or glowing fruit somewhere on the enclosure and you can install ESXi and virtualize OSX 
|
|
#
?
Oct 8, 2013 19:34
|
|
|
It's on my ABC affiliate right now!
|
|
#
?
Oct 8, 2013 20:24
|
|
|
I am still surprised every day by how many "vmware engineers" still will not virtualize vcenter, or do not know how to recover from an environment down issue with a virtual vcenter. "Well what happens if the host goes down that vcenter is on?" HA carries it to a new host "Well what happens if your environment powers down?" You can attach to the host and force start vcenter "Well what happens if it is 50 or 100 hosts" If you have 100 hosts you should be carving out a management cluster which would mitigate quite a bit of "guessing" if you don't want to write a script to query hosts for the vm labeled your vcenter.
|
|
#
?
Oct 8, 2013 22:20
|
|
|
Dilbert As gently caress posted:I am still surprised every day by how many "vmware engineers" still will not virtualize vcenter, or do not know how to recover from an environment down issue with a virtual vcenter. It started with people not wanting to virtualize SQL servers, then not wanting to virtualize domain controllers. Virtualizing vCenter will be the last domino to topple.
|
|
#
?
Oct 8, 2013 22:28
|
|
|
Don't really need a management cluster, just a DRS rule. Mine's virtualized and pinned to the first two hosts in the cluster. That said, I've had two complete power outages and had cause to regret virtualzing VCenter both times: once due to HA powering up the entire environment on the first host that booted, and the other due to a really lovely HA bug in 5.0 that broke HA for any machine that had been storage vmotioned on a VDS. That bug really represents that this stuff is far from bullet proof. Good IT guys are risk averse, and from my experiences over the last few years, virtualizing vcenter can easily add more risk than it removes. After all, big operations tend to have mature strategies in place to protect physical servers.
|
|
#
?
Oct 8, 2013 22:36
|
|
|
Any good resources to provide for a manager who's wary of virtualizing SQL? We're running HyperV 2012 but he's really skittish SQL VMs.
|
|
#
?
Oct 8, 2013 23:18
|
|
|
Erwin posted:It sounds like he only has one host at DR. If that's the case, he's not losing anything (but a bill for SAN hardware). Never thought of using Veeam. Nice idea. Yeah we only have the one host at our second site. We'd be replicating just the mission critical stuff like mail and database. We cant afford to run duplicate hardware at both sites so we'd have to choose what to spin up. Veeam has an option for 'Continuous' replication. Is there any reason not to use this option if our WAN can handle it?
|
|
#
?
Oct 8, 2013 23:21
|
|
|
Swink posted:Veeam has an option for 'Continuous' replication. Is there any reason not to use this option if our WAN can handle it? You spot corruption on VMFS volume A which corrupted VMA, VMB, and VMC. Guess what happens to your replicas? Crackbone posted:Any good resources to provide for a manager who's wary of virtualizing SQL? We're running HyperV 2012 but he's really skittish SQL VMs. You can try this book, looks like it's the new revision http://www.amazon.com/Virtualizing-...irtualizing+sql KS posted:Don't really need a management cluster, just a DRS rule. Mine's virtualized and pinned to the first two hosts in the cluster. I agree with you that you should mitigate and minimize risks whenever possible, it really comes down to the person implementing the solution. I would rather have someone buy a box for vcenter that doesn't know how to recover from a all hosts down issue, or failed HA vcenter start. To me, a virtual vcenter has more benefits that outweigh the risk. Dilbert As FUCK fucked around with this message at 23:33 on Oct 8, 2013 |
|
#
?
Oct 8, 2013 23:28
|
|
|
blah forget SA does not have automerge
|
|
#
?
Oct 8, 2013 23:29
|
|
|
Dilbert As gently caress posted:You spot corruption on VMFS volume A which corrupted VMA, VMB, and VMC. Guess what happens to your replicas? Whats the difference between that happening on a continuous cycle or at 15 min intervals?
|
|
#
?
Oct 8, 2013 23:50
|
|
|
Swink posted:Whats the difference between that happening on a continuous cycle or at 15 min intervals? About 15 minutes worth of calls. Nah it comes down to your RTO/RPO on what you should do, most of my clients have a 2 or 4hr window for replication jobs which generally gives us enough time for a client to realize something is corrupt and we should fail over; and enough buffer to stop a replication if VMA/VMB become, corrupt to stop a replication. There also may be a hit on storage and WAN with continuous. However, it really depends on the use case and requirements of the environment
|
|
#
?
Oct 8, 2013 23:54
|
|
|
Okay so I randomly PM a low number VCDX about poo poo which to become a VCDX(and I know he hates the poo poo out of me; and is which is why I want to repay him); And I am sure annoys the piss out of him. So I want to ask you all: I really want to be a VCDX one day; I'll go through some poo poo to get there but I want to be a VCDX/CCNP/EMISCA when I am 25-26.... What do you think best practice could help me on? I mean poo poo any advice is useful I want to make best of it! I WANT TO BE A VCDX! Dilbert As FUCK fucked around with this message at 03:48 on Oct 9, 2013 |
|
#
?
Oct 9, 2013 03:40
|
|
|
Dilbert As gently caress posted:Okay so I randomly PM a low number VCDX about poo poo which to become a VCDX(and I know he hates the poo poo out of me; and is which is why I want to repay him); And I am sure annoys the piss out of him. So I want to ask you all: Are you drinking again? That said I'm not sure I understand your question but if it's something you really want then you should prepare for a long hard road. Once you get the two exams down you're going to need to start building some practical experience doing design work. When you're doing your design if the only justification for your decisions are "it's a best practice" then you're not really going to get very far. You should be able to speak with some level of depth on everything your design touches. You'll need to be able to speak with clarity and confidence in everything you do. So how do you prepare? Get involved and lead design projects. Be able to speak to network and storage people in their language. Have a basic understanding of common applications (things like SQL server, Exchange, etc. Color this based on what your customer base/industry is doing) Start to get an understanding of the business side of things (the whys of what you're doing.) This becomes crucially important when you start taking DR, service levels, etc. It's worth getting an understanding of things like TOGAF and/or Zachmann to give you some ideas for approach here. Develop your communication skills! This is crucially important as an architecture in your head is completely worthless if you can't properly communicate it to your customers/team/stakeholders, etc. You'd posted a visio earlier (I think in this thread) that I wanted to ask questions on but got sidetracked with work. Some high level feedback was that I was unclear what it was you were trying to tell me. At it's face value I saw things in it that didn't make a lot of sense (hosts connected to inside and DMZ networks for example) but they might have been "dotted line" connections such as an open port through the firewall. You should know how to talk to technology people so you can develop your architecture but you need to know how to talk to business people to find out things like your business requirements, drivers, etc. Also to communicate to them the benefits of your design and how it meets any requirements/addresses risks that matter to them. All that said I don't hate anyone! I just get busy.
|
|
#
?
Oct 9, 2013 06:01
|
|
|
Dilbert As gently caress posted:... a low number VCDX ... Appreciate you making this important distinction, as we all know only the first 80 or so were worth a drat. Edit: I guess I can try to be helpful, too. Keep your eyes open. Soak up all you can about the storage, network, compute, and DR, as these are all things beyond just virtualization that you'll need to detail in your design, and will need to be able to speak with authority on in your defense. Read the VCDX blueprint and application--they outline in very black and white terms exactly what level of effort is expected in the design submission. As you feel like you're getting close to being able to submit, start reading the VCDX Bootcamp book, watching the VCDX vBrownBag sessions, and viewing VMware's VCDX Bootcamp videos to get a better idea of what the defense process feels like. Get as many mock panels under your belt as possible, and make sure they include the design and troubleshooting scenarios. Pantology fucked around with this message at 15:28 on Oct 9, 2013 |
|
#
?
Oct 9, 2013 13:09
|
|
|
So I posted this in the IT Cert thread, but I'll post it here as well: So heads up, if you're from Alaska, Arkansas, Kentucky, or Minnesota, you can't take the SCC Vmware class. I just got this email. quote:State Authorization Compliance: So any other online VCP classes that don't cost an arm and a leg? I got approval for my job to pay for this one, so even if it's 2 or 3 times the cost of the Stanly class, I think I'd still be OK.
|
|
#
?
Oct 9, 2013 20:15
|
|
|
Dilbert As gently caress posted:You spot corruption on VMFS volume A which corrupted VMA, VMB, and VMC. Guess what happens to your replicas? The bigger problem with continuous replication or anything on a short schedule is, depending on change rate, storage performance, and planetary alignment, the stun length when deleting snapshots. It can be long enough to drop a few pings, or even longer. You'll have users at your desk with pitchforks. Every 15 minutes. In my mind, Veeam replication is better used at night, say once per night. Combine that with SQL mirroring/log shipping and DFS-R, and intelligently lay out your changes for a low RPO. Of course I'm not protecting a mail server, just internally developed software, so we can silo our writes to SQL and DFS shares as much as we want.
|
|
#
?
Oct 9, 2013 21:18
|
|
|
Alright, what did I break? (Well, I know what I broke, just not why and how to fix yet.) We were at a 2003 Functional level for AD and Forest. We raised to 2012 Level for both. Now SSO is broken. http://blog.uvm.edu/jgm/2012/10/31/vsphere-51-train-wreck/ That seems to suggest that I might have to make sure the SSO service is running under a domain account a give that account permissions to certain objects in AD, but I'm just wondering if anyone else has seen this before. Edit: Really? SSO doesn't query DNS for the AD controllers and hard codes them in when the service was installed? That would explain it since we retired all our old 2003 domain controllers at the same time we did the domain level upgrade. Oh well, at least I know how to fix this now. bull3964 fucked around with this message at 20:02 on Oct 10, 2013 |
|
#
?
Oct 10, 2013 19:26
|
|
|
bull3964 posted:Really? SSO doesn't query DNS for the AD controllers and hard codes them in when the service was installed? That would explain it since we retired all our old 2003 domain controllers at the same time we did the domain level upgrade.
|
|
#
?
Oct 10, 2013 22:34
|
|
|
Dilbert As gently caress posted:I mean poo poo any advice is useful I want to make best of it! I WANT TO BE A VCDX! Honestly, go work for a partner or VAR. I thought I was going to get a VCDX out of my current job's brand new meganetwork I am going to be doing a lot of implementation on, but they won't let me be involved in the project planning or procurement process which is all happening at the VP level. They let me full upgrade our vBlock 300 though which was a very nice boost to my VCAP-DCA plans. If you're at a VAR or partner, you'll get thrown into network design tasks every day and eventually you take lead on one and have the perfect VCDX project. Helps to have a VCDX mentor there too. bull3964 posted:Really? SSO doesn't query DNS for the AD controllers and hard codes them in when the service was installed? That would explain it since we retired all our old 2003 domain controllers at the same time we did the domain level upgrade. VUM IPs are hardcoded into vCenter too, so if you retire a vCenter I think you have to go into ADSI edit to fix it. Edit: Two new interesting VMUG Webcasts: http://www.vmug.com/p/cm/ld/fid=38&source=5 One on performance tuning and one on upgrading to 5.5, both hour long lunch things in the next six weeks. (And an EMC backup event I'm sure they're going to pitch Data Domain / Avamar the whole time) El_Matarife fucked around with this message at 22:58 on Oct 10, 2013 |
|
#
?
Oct 10, 2013 22:48
|
|
|
bull3964 posted:Edit: Don't hardcode your domain controllers. You should have an A record for your domain, use that instead.
|
|
#
?
Oct 10, 2013 23:17
|
|
|
bull3964 posted:Really? SSO doesn't query DNS for the AD controllers and hard codes them in when the service was installed? That would explain it since we retired all our old 2003 domain controllers at the same time we did the domain level upgrade. VMware 
|
|
#
?
Oct 10, 2013 23:18
|
|
|
FISHMANPET posted:Don't hardcode your domain controllers. You should have an A record for your domain, use that instead.
|
|
#
?
Oct 10, 2013 23:20
|
|
|
FISHMANPET posted:Don't hardcode your domain controllers. You should have an A record for your domain, use that instead. I didn't hardcode anything. When I installed SSO, it queried DNS and picked two domain controllers at random and put them in as primary and secondary in Identity Sources in SSO. When I retired those two domain controllers, it broke SSO since it had nothing left to query against. My question is, why in the hell isn't SSO just querying _ldap._tcp.dc._msdcs.Domain_Name for the domain controllers instead of actually putting a primary and secondary LDAP server in Identity Sources?
|
|
#
?
Oct 10, 2013 23:31
|
|
|
When I replaced my two domain controllers I just said 'gently caress it' and created a new authentication source with ldap:\\ad.domain.blahblah and that's worked just fine. The problem is that it's not really treating it like an AD source, it's just treating it like a dumb LDAP server and doesn't look for any SRV records. Yeah it's colossally stupid, but if you just put in the A record of your domain it will work, as long as that record always points to at least one active domain controller (which, if you run Microsoft DNS will all happen automatically).
|
|
#
?
Oct 11, 2013 00:23
|
|
|
El_Matarife posted:Honestly, go work for a partner or VAR. I thought I was going to get a VCDX out of my current job's brand new meganetwork I am going to be doing a lot of implementation on, but they won't let me be involved in the project planning or procurement process which is all happening at the VP level. They let me full upgrade our vBlock 300 though which was a very nice boost to my VCAP-DCA plans. If you're at a VAR or partner, you'll get thrown into network design tasks every day and eventually you take lead on one and have the perfect VCDX project. Helps to have a VCDX mentor there too. Yeah my next job is going back to working with a VAR. I would have loved my last job if I could have focused on external clients more than being a puppet spinning a thousand plates. I work with a VMware partner but I'm not too happy as 40% of my job is working with end users and SMB of SBS servers. My current design is basically a reversed engineered VATSEE(which I didn't know existed till my DCD class), and basically virtualizing a whole college at some point. I mean I have the design, the networks, storage, and security all drawn out in a completely scalable solution. Hell I do these designs for fun because I get bored. I really need to post my complete work but honestly feel it would be better with a live audience. Eh maybe I can use this to populate my neglected website. I basically incudes, Horizon, vApp'd Servers, low level vButt, and vSphere replication. As well as a complete redesign of network/storage/compute. 1000101 posted:Words and solid advice Actually yeah I was, been trying to be more sociable with my coworkers and all wanted to drink/go out that night, my bad. I appreciate the input. Dilbert As FUCK fucked around with this message at 03:12 on Oct 12, 2013 |
|
#
?
Oct 12, 2013 02:31
|
|
|
 HUZZAH 4 Gen8's 384GB/16x2proc/dual FC8Gb/s/8x1Gb/e up and ready to roll THE DESIGN/IMPLEMENT FUN HATH BEEN DOUBLED! Can't wait too fully implement now the HW is in place. note:not my cable work, come december it will be nicer Dilbert As FUCK fucked around with this message at 06:00 on Oct 12, 2013 |
|
#
?
Oct 12, 2013 05:45
|
|
|
I have recently started reading about VMware and ESXi and the wondrous things it can do with the right hardware, and naturally I have a question: If I run two instances, one with Win 7 and one with Debian (for LAMP and whatnot), would it then be possible to run the Win 7 instance as a normal HTPC hooked up to my TV? Thanks goons!
|
|
#
?
Oct 14, 2013 16:54
|
|
|
It's POSSIBLE with VMDirectPath, but I don't think anyone will say that it's a good way to run it. Your better bet would be to just run Win7 as the HTPC with VMware Workstation running your Debian VM in the background.
|
|
#
?
Oct 14, 2013 17:02
|
|
|
I was guessting that I would have to go through some none delicate way to do it, thanks tho!
|
|
#
?
Oct 14, 2013 17:51
|
|
|
My SSO 5.5 installation rolls back and the fixes here don't make a difference. The 5.1 SSO install it's upgrading is entirely vanilla, self signed certs, default everything. Blog comments on posts for this issue suggest that the KB article isn't helping for a bunch of people. Support is more interested in closing the ticket than actually fixing the problem, and they want me to reinstall VCenter to fix it. Just wondering if anyone has had better luck from them? It seems they simply cannot do an SSO release without loving up.
|
|
#
?
Oct 14, 2013 19:54
|
|
|
evol262 posted:They sell Minis with Server for business that want to be Mac-y, and they do a lot better in that market. Yeah, and it blows. I wish I could toss this useless box and run it on one of our many real servers.
|
|
#
?
Oct 16, 2013 09:25
|
|
|
HalloKitty posted:Yeah, and it blows. I wish I could toss this useless box and run it on one of our many real servers. That's an entirely different argument. The Xserves and XSANs were also unmaintained pieces of poo poo that got refreshes less often than Mac Pros with questionable uptake in the enterprise.
|
|
#
?
Oct 16, 2013 16:10
|
|
|
I'm shocked Apple won't just team up with Oracle and announce OSX for SUN x86 servers at ~$5,000 a socket. It'd take care of the business / enterprise space and Jobs was really, really close friends with Larry Ellison. So I apparently managed to install both my vCenter SSOs as primaries when doing my 5.1 upgrade, I'm going to have to completely uninstall and reinstall the DR site SSO. Total pain, I can't believe I missed that checkbox. The error message in the JoinTool log is spectacularly worthess too: Operation "Join instance XXXXX" failed: : Action: Join Instance Action: Prepare for Join Problem: Remote VC's XXXXXX SSO domain ID is does not match."
|
|
#
?
Oct 16, 2013 20:11
|
|
|
El_Matarife posted:I'm shocked Apple won't just team up with Oracle and announce OSX for SUN x86 servers at ~$5,000 a socket. It'd take care of the business / enterprise space and Jobs was really, really close friends with Larry Ellison. There's just not a reason. If you want "Certified UNIX" $5k servers, get AIX, HP-UX, or Solaris/SPARC. OSX has zero advantages over Linux/BSD, and a Mini is fine for your "I need to remotely wipe your iPhone if you leave the company" needs.
|
|
#
?
Oct 16, 2013 22:08
|
|
|
evol262 posted:It didn't get uptake because OSX Server is a POS that you only really need these days for managing iOS devices, and for nothing in 2005, unless you want to run really expensive servers on an OS that's slower at every server task than Linux and whose major driver is a UI that doesn't matter on a server chassis, and...
|
|
#
?
Oct 17, 2013 04:11
|
|
|
Yeah, the point is not to virtualize OSX Server. The point is to have tons of copies of OSX running on a beefy ESX box. I imagine mostly for VDI. A ton of places use OSX for OpenGL work, and now the VDI is getting into graphics work, OSX would be useful.
|
|
#
?
Oct 17, 2013 05:22
|
|
|
I have a strange question. I have a client that is needing to buy a new server. The issue is that the application he needs to run on it will only run under 32-bit windows which for a server seems like a bad idea, at least from a RAM standpoint. So I'm suggesting to him that create a virtual 32 bit system on his new 64-bit box either using Virtualbox or probably VMware. He has a concern that the VM won't be very responsive, so I have suggested that maybe he invest in a semi-large SSD drive and run the VM off of that and have a separate traditional drive for whatever data archiving he may need to do. I'm also suggesting that he get as much RAM as the server can handle within reason, it's only going to be running a small tv station traffic system not doing GIS. The company that makes the application recommends an i5 with 4 gigs of RAM. I'm thinking getting a i7 would be the best along with a 256 gig or bigger SSD drive. Is this a feasible idea or should I look in a different direction?
|
|
#
?
Oct 17, 2013 05:53
|
|
|
diremonk posted:I have a strange question. I have a client that is needing to buy a new server. The issue is that the application he needs to run on it will only run under 32-bit windows which for a server seems like a bad idea, at least from a RAM standpoint. So I'm suggesting to him that create a virtual 32 bit system on his new 64-bit box either using Virtualbox or probably VMware. That is pretty much what virtualization was meant for, but it sounds like you may be in over your head. You definitely don't want to use Virtualbox. You want ESXi, HyperV, or XenServer. And if you're talking about buying a new server, it should not have a Core i series chip in it.
|
|
#
?
Oct 17, 2013 12:34
|
|
|
|
| # ? Apr 29, 2024 02:36 |
|
|
Misogynist posted:Something datacenter-grade (hell, even just form factor) that you can virtualize OS X is critical. This is not arguable. There are tons of Mac development shops that slave their continuous integration onto a stack of Mac Mini servers stuffed into a colo rack. Nobody likes doing this.
|
|
#
?
Oct 17, 2013 12:56
|
|