|
Martytoof posted:Anyone who is planning on taking the Stanly.edu VCP-DCV course, I urge you to get acquainted with the clock that drives their servers: Are you able to use a home lab if you have one? Also I'll have to setup vCheck. Never heard of it before.
|
# ? Sep 7, 2013 01:07 |
|
|
# ? May 28, 2024 15:54 |
|
evil_bunnY posted:Did you really have to ruin it for everyone? To be fair it was a pretty dumb idea. I wasn't brain fired lurking the forums on lunch it probably would have been a better response, but yeah gotta love it...
|
# ? Sep 7, 2013 02:31 |
|
El_Matarife posted:
I was actually informed that this was not true by a VCAP-DCD at lunch today a recent VCDX had a largely fictional design that passed.
|
# ? Sep 7, 2013 02:52 |
|
Dilbert As gently caress posted:I was actually informed that this was not true by a VCAP-DCD at lunch today a recent VCDX had a largely fictional design that passed. You can just say you heard it from a guy. We don't need his resume.
|
# ? Sep 7, 2013 03:34 |
|
Pantology posted:You can just say you heard it from a guy. We don't need his resume. He's one of the guys I really respect so I thought I might include a cert level as appose to a "I heard from the wind"
|
# ? Sep 7, 2013 04:10 |
|
Moey posted:Are you able to use a home lab if you have one? Nope.
|
# ? Sep 7, 2013 04:24 |
|
Thinking about doing some vmugs or vSAMugs on Wednesday night anyone interested? Might as well join the VCAP's/VCDX's of SA' and help out.
|
# ? Sep 7, 2013 05:53 |
|
Erwin posted:Don't do this. Unless you can't afford another Windows license for some reason, you should just have a DC that's just a DC (plus DNS and DHCP or whatever). Maybe you did it so you didn't have to buy more hardware because they're all physical, but one of the benefits of virtualization is that there's no reason not to silo things off. I think long term the DR data will be moved to another site so then the DR server will 'just' be a DC, I think we have to have it in place while some governance goes on but I will find out. Budget is tight so I don't think I will get any more servers/licenses. Where there's a will there's a way...
|
# ? Sep 7, 2013 13:14 |
|
Misogynist posted:Eh? The default maximum clock skew in MIT Kerberos is 5 minutes, which is also considered the recommended best practice for Microsoft's "Maximum tolerance for computer clock synchronization" policy setting. This is also the threshold where some SSL sites will begin to give you problems. Learn something every day. I'd always assumed that it was much less than 5 minutes.
|
# ? Sep 7, 2013 20:33 |
|
Oh what VDP supports replication to an avamar device in 5.5. How did I miss that?
|
# ? Sep 7, 2013 23:50 |
|
So I found a side benefit to the VCA. I haven't talked about it at work (other than to one guy on another shift who asked me about it and has been borrowing my Mastering Vsphere book) but I printed out the certificate and pinned it to my cubicle. I think it might be an excellent way to identify "Expert Beginners" because a "competent" person couldn't possibly feel threatened by that sheet of paper. To an expert beginner it might as well be radioactive.
|
# ? Sep 8, 2013 03:18 |
|
Dr. Arbitrary posted:So I found a side benefit to the VCA. I haven't talked about it at work (other than to one guy on another shift who asked me about it and has been borrowing my Mastering Vsphere book) but I printed out the certificate and pinned it to my cubicle. Yeah I am recommending it to most of the ICM students I help out. To my friend it may be a "super VSTP" but to others who are starting it may prove very useful. I feel it is a good exam overall, nothing amazing but definately a good starting point.
|
# ? Sep 8, 2013 04:22 |
|
Dilbert As gently caress posted:Thinking about doing some vmugs or vSAMugs on Wednesday night anyone interested? Are you thinking Google Hangouts or what? I don't suppose someone has a Webex / GoTo Meeting account from work we can borrow?
|
# ? Sep 8, 2013 19:54 |
|
Given licensing is a non-issue, is there any inherent advantage to running VMs directly on hyper-v server vs ws2012r2?
|
# ? Sep 10, 2013 05:56 |
|
Hadlock posted:Given licensing is a non-issue, is there any inherent advantage to running VMs directly on hyper-v server vs ws2012r2? No, the only real considerations are cost,(Hyper-V server is free, WS2012R2 costs money for every 2 physical procs), command line vs GUI management(Hyper-V Server is command line only), Licensing(Hyper-V server provides no in-box licensing, standard server has 2 VM servers allowed per license, datacenter provides unlimited server VMs) Basically what I tell folks, is if you are virtualizing non-MS OSes, or doing VDI, Hyper-V server is cheaper and worth a look. If doing server virtualization with MS OSes, datacenter is usually the best bet if you are virtualizing approximately 8-10 VMs for every 2 physical processors.
|
# ? Sep 10, 2013 16:21 |
|
Hadlock posted:Given licensing is a non-issue, is there any inherent advantage to running VMs directly on hyper-v server vs ws2012r2? Hyper-V Server has less Windows roles, so less patching, which means more uptime.
|
# ? Sep 11, 2013 11:16 |
|
Mully Clown posted:Hyper-V Server has less Windows roles, so less patching, which means more uptime. actually patching is pretty much the same if you setup Windows Server 2012 Core with hyper-v and only hyper-v.
|
# ? Sep 11, 2013 14:34 |
|
http://www.datacenterknowledge.com/archives/2013/09/11/cisco-acquires-whiptail-for-415-million/ Sounds like Cisco is a believer in distributed storage on compute. If vSAN and CloudStack Swift / Cinder are the future, the big monolithic SAN guys should be worried. I always figured Cisco would buy NetApp or another mid-tier SAN vendor to complete their compute / network / storage stack. This also dovetails nicely with the read cache support in VMware 5.5, though I suspect this might be at the chassis / FEX level rather than individual blades. Or possibly they want to build a Dell VRTX style chassis with a "SAN" integrated as a blade. El_Matarife fucked around with this message at 18:43 on Sep 11, 2013 |
# ? Sep 11, 2013 18:26 |
|
Sitting in the testing center waiting to go in to my VCP exam. I'll know if all this studying worked in about 2 hours
|
# ? Sep 12, 2013 14:57 |
|
Syano posted:Sitting in the testing center waiting to go in to my VCP exam. I'll know if all this studying worked in about 2 hours You got that poo poo man, no worries.
|
# ? Sep 12, 2013 15:09 |
|
Killed it with a 450!
|
# ? Sep 12, 2013 16:29 |
|
Congrats!
|
# ? Sep 12, 2013 16:34 |
|
Syano posted:Killed it with a 450! Nice work dude, that's a great score.
|
# ? Sep 12, 2013 19:54 |
|
Syano posted:Killed it with a 450! Nice job! vCenter installs tonight for my class, hope the environment/vAPPs holds nothing like 26 installs of vCenter 5.1 GA running against a FAS 2240-4. Luckily the Vapp has the SQL and ODBC connections in place so that takes out a good portion.
|
# ? Sep 12, 2013 21:10 |
|
Data corruption is always a good bug! Server 2012 with E1000E NIC. http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=2058692
|
# ? Sep 13, 2013 22:12 |
|
vmxnet3 is the way for me
|
# ? Sep 13, 2013 22:29 |
|
Syano posted:vmxnet3 is the way for me Same here, just giving a heads up. If anyone is unsure, a quick way to check your entire environment is use RVtools, go to the vNetwork tab, then sort adapter.
|
# ? Sep 13, 2013 22:36 |
|
Alright gently caress VATSE I can do it easier and cheaper Need 2 more goons to test this poo poo let me know
|
# ? Sep 15, 2013 02:05 |
|
Hadlock posted:So basically, HyperV -> WS2012R2 VM -> Physical Storage Space -> VM running on Hyper-V Update, this actually works, since you can bind the physical drives to the VM, and then create the mirrored storage pool inside the VM, and associated volume(s). I'm able to saturate my physical 1GBE NIC @ 984Mbps sustained using consumer grade hardware, so I'm happy with whatever latency there is. Hadlock fucked around with this message at 08:24 on Sep 15, 2013 |
# ? Sep 15, 2013 08:20 |
|
So, I think I may have finally convinced my boss that we need to start virtualizing before we're seriously left behind. We've never virtualized before, mainly due to a zero dollar budget. Anyway, for our first server, I'm thinking about a simple 2 cpu poweredge from Dell with 128GB of memory. I was going to do two 10k 300GB drives in RAID-1 for the hypervisor (VMware vSphere ESXi). I know this is not a huge powerful server but I need to keep the upfront cost reasonable or my boss will shut it down. Then I was going to purchase either a DAS or NAS (iSCSI) for the datastore, but my question comes down to how to RAID the drives in the datastore. If I get 16x 10k RPM 600GB drives, what is the best way to RAID them for A) the best reliability/redundancy while B) not giving up too much space? I was thinking like a one large RAID-60 scenario with 2x 8 drive groups, but a friend with more experience and a better job told me to do 4 groups of RAID-10. This would mean I'd have to more carefully balance my VMs over the drives so that one set isn't getting hammered and the other is idle. What would you guys do or what is best practice?
|
# ? Sep 16, 2013 00:42 |
|
IT Guy posted:So, I think I may have finally convinced my boss that we need to start virtualizing before we're seriously left behind. We've never virtualized before, mainly due to a zero dollar budget.
|
# ? Sep 16, 2013 00:55 |
|
Misogynist posted:Before going down this route, let's make sure we've got your priorities in order, because your plan to buy one single server to virtualize a bunch of things worries me. The best way to proceed forward is to make sure sure your budget situation is straight and your requirements are realistic. If you haven't properly considered the other implications of virtualizing, your budget for your main server will probably end up being less than you originally intended. We really only have 3 mission critical things, the two domain controllers (which I would never but on the same hardware), the file server which is already on new hardware and won't be virtualized, and then a new Exchange server that will be going on this hardware. So here is my plan, Exchange server and spam/filtering server both virtualized, because if one or the other goes down, well they both are down pretty much. Then I'd be throwing the secondary DC on it, as well as a web server. So that's 4 servers. I haven't really thought of anything else to put on it. Then if that all goes smoothly, we'd purchase another identical server and virtualize the rest of it on that one. We have 13 servers in total so we could get away with 2-3 physical machines easy. As for backups, we wouldn't be backing up the VM to start, we'd be using our backup exec licenses on the actual windows install and backing up traditionally, at least for the first little bit. Once we have things virtualized and no longer on traditional hardware, we'd look into something like Veeam to backup the VM. Or... should I just be looking for a new job? edit, actually I forgot we have the POS/Inventory server which we'll be upgrading in 2014 which could possibly be virtualized as well. IT Guy fucked around with this message at 01:32 on Sep 16, 2013 |
# ? Sep 16, 2013 01:26 |
|
IT Guy posted:We really only have 3 mission critical things, the two domain controllers (which I would never but on the same hardware), the file server which is already on new hardware and won't be virtualized, and then a new Exchange server that will be going on this hardware. So here is my plan, Exchange server and spam/filtering server both virtualized, because if one or the other goes down, well they both are down pretty much. Then I'd be throwing the secondary DC on it, as well as a web server. So that's 4 servers. I haven't really thought of anything else to put on it. Then if that all goes smoothly, we'd purchase another identical server and virtualize the rest of it on that one. We have 13 servers in total so we could get away with 2-3 physical machines easy. As for backups, we wouldn't be backing up the VM to start, we'd be using our backup exec licenses on the actual windows install and backing up traditionally, at least for the first little bit. Once we have things virtualized and no longer on traditional hardware, we'd look into something like Veeam to backup the VM. what budget are you working with?
|
# ? Sep 16, 2013 01:39 |
|
Dilbert As gently caress posted:what budget are you working with? Probably about 20k. But I never get a budget, it's literally, propose something, they think about it for a couple months and then something happens.
|
# ? Sep 16, 2013 01:59 |
|
Just going to jump back a bitIT Guy posted:Anyway, for our first server, I'm thinking about a simple 2 cpu poweredge from Dell with 128GB of memory. I was going to do two 10k 300GB drives in RAID-1 for the hypervisor (VMware vSphere ESXi). I know this is not a huge powerful server but I need to keep the upfront cost reasonable or my boss will shut it down. Depending on what your applications need 2 quad cores and 128GB can be a LOT. Tools like Dell's DPACK and other tools like capacity planner can help show what you actually need in terms of Network/Storage/CPU/Ram, run this tool unless you have some monitoring tools which can report those numbers too you. Also ESXi really does not need Raid 1, it runs in ram and can be installed on a thumb drive. quote:Then I was going to purchase either a DAS or NAS (iSCSI) for the datastore, but my question comes down to how to RAID the drives in the datastore. If I get 16x 10k RPM 600GB drives, what is the best way to RAID them for A) the best reliability/redundancy while B) not giving up too much space? Completely depends on what the applications needs are and the data that resides on those stores. While an exchange or FS might be able to sit happy on a raid 6 array of 2TB LN-SAS drives, things such as SQL DB's may need RAID 10 Design your storage based on the requirements of the systems not by Size. IT Guy posted:We really only have 3 mission critical things, the two domain controllers (which I would never but on the same hardware), the file server which is already on new hardware and won't be virtualized, and then a new Exchange server that will be going on this hardware. So here is my plan, Exchange server and spam/filtering server both virtualized, because if one or the other goes down, well they both are down pretty much. Would a P2V of the Fileserver to re-use the HW be out of the question? frankly if you have a nice shiny new server why not use it or put a plan in place to migrate it to virtual and re-purpose the host for failover or additional resources? Something like essentials or Essentials + can span 3 hosts, and give you some more flexibility when paired with a NAS. quote:Then I'd be throwing the secondary DC on it, as well as a web server. So that's 4 servers. I haven't really thought of anything else to put on it. Then if that all goes smoothly, we'd purchase another identical server and virtualize the rest of it on that one. We have 13 servers in total so we could get away with 2-3 physical machines easy. As for backups, we wouldn't be backing up the VM to start, we'd be using our backup exec licenses on the actual windows install and backing up traditionally, at least for the first little bit. Once we have things virtualized and no longer on traditional hardware, we'd look into something like Veeam to backup the VM. I'm a bit confused here, are you trying to stage the projects purchasing? Or are you buying a set amount upfront to handle HA, backups, and operations; or buying one and then hoping management deems another one to finish the project is okay.
|
# ? Sep 16, 2013 02:24 |
|
Dilbert As gently caress posted:
The latter. What's happening is that we upgrade a couple servers a year, and everytime we do, we purchase new hardware for a standalone machine. This causes problems where if we need a new server for something, it gets denied for hardware costs and we end up with servers that are a DC, an SQL box, a web server, application server and a bunch of other random poo poo thrown on it which makes it a nightmare. If we could simply purchase another Windows server license and startup a new VM, that would be so much better. You have to keep in mind, we're barely a mid-size business so all this enterprise poo poo is beyond them. IT Guy fucked around with this message at 02:40 on Sep 16, 2013 |
# ? Sep 16, 2013 02:36 |
|
IT Guy posted:You have to keep in mind, we're barely a mid-size business so all this enterprise poo poo is beyond them.
|
# ? Sep 16, 2013 02:51 |
|
adorai posted:Even small business IT can be run like an enterprise these days, often times to significant cost savings. I think you really need to do a cost savings analysis to show your management how much you could have saved if you had done this say, three years ago. Interesting point. I think you're right, and now that I think about it, this is what I'll do.
|
# ? Sep 16, 2013 02:59 |
|
Dilbert As gently caress posted:Depending on what your applications need 2 quad cores and 128GB can be a LOT. Tools like Dell's DPACK and other tools like capacity planner can help show what you actually need in terms of Network/Storage/CPU/Ram, run this tool unless you have some monitoring tools which can report those numbers too you. Also ESXi really does not need Raid 1, it runs in ram and can be installed on a thumb drive. Besides this, consider just running the hypervisor off of an embedded SD card, which Dell offers on most servers. You'll have increased reliability without the spinning drives, and it'll shave a decent percentage off the price to boot. Plow the savings into two quad-port Intel NICs (gently caress Broadcom forever) so that you can have dedicated interfaces for iSCSI, vMotion, VM networks and management with at least 2 ports dedicated to each function for redundancy. Edit: I also strongly agree with Misogynist that with just one host you are really getting minimal benefit from virtualization, and potentially making your environment worse. Essentials Plus kit on 2 hosts (each host beefy enough to run all of your VM's on its own for redundancy) is kind of the minimum I'd ever recommend for a production environment, even for a small business. Without that you have no HA, can't vMotion things around to do patching or hardware swaps, and have a glaring single point of failure. You've consolidated a bunch of lovely old servers into one, which is good, but at great risk to stability. Docjowles fucked around with this message at 17:08 on Sep 16, 2013 |
# ? Sep 16, 2013 17:01 |
|
|
# ? May 28, 2024 15:54 |
|
So I haven't really looked past vSphere ESXi, but two of you mentioned Essentials so I went and looked at what it was and apparently it's cheap as hell and comes with vSphere Server? That means I can migrate VMs from one host/datastore to another, correct? When I was looking into it, I was looking at standard and saw that it was about 10k. Am I wrong here, is essentials really only $500 and is licensed for 3 hosts with 2cpu each? edit: oh ok, Essentials + is $5k (still reasonable) and comes with HA and vMotion. IT Guy fucked around with this message at 18:04 on Sep 16, 2013 |
# ? Sep 16, 2013 18:01 |