|
Wow actually starting to like citrix in some respect the more I deploy it vs other engineers. It really makes me wonder what others are doing with those X hours scheduled.... Probably should have factor'ed in the typical "next until I hit finish" factor after: HEY Dilbert I hit next until Symantec vShield(or N&S) was done why does my environment suck?
|
#
?
Nov 9, 2013 06:00
|
|
|
|
| # ? May 8, 2024 06:27 |
|
|
Anyone know much about UCS? I'm hearing rumors about it at work. I've tried to do some research but I'm not savvy enough yet to really understand. How does it relate to virtualization?
|
|
#
?
Nov 9, 2013 06:17
|
|
|
Dr. Arbitrary posted:Anyone know much about UCS? Worked/sold with UCS for a good while at my last job. What do you want to know about them? My opinion? Cons's They are overpriced compared to any competitor, basically whitebox HP's with a custom java management console. Unless you an a var you where you are a full cisco partner and get mega BOGO's or discounts they are uber expensive to most end customers They boast "ram and CPU configs" you can get "nowhere else" because you can get it better else where if you go to hp's sce or dell since 2010 Shipping SUCKS rear end, dell can send it to my in 4 days why does a PCI card delay my order by 21 days? gently caress YOU Pro's Converged networking is cool, and for healthcare or x-large environments. If you have a customer needing to fill X budget or they get less next year, while maximizing profit; UCS is best If you are an all cisco shop server/network/VIC's there are some pretty sweet tools out there for you I'm sure some goons can point out more but honestly I don't see much to UCS. Dilbert As FUCK fucked around with this message at 06:28 on Nov 9, 2013 |
|
#
?
Nov 9, 2013 06:25
|
|
|
I think you gave me exactly what I was looking for. I think we've already made a purchase but maybe I can talk to the managers at work and learn a little bit about how we decide what to buy.
|
|
#
?
Nov 9, 2013 08:07
|
|
|
Dilbert As gently caress posted:Worked/sold with UCS for a good while at my last job. What do you want to know about them? Cisco's cutting some pretty deep discounts for them (smartplay bundles get you in the door for dirt cheap). It's pretty price competitive against HP C7000 and featurewise I'd put it ahead. It's pretty much the farthest thing from a whitebox HP with "a custom java management console." UCSM is pretty damned extensible and we've done some pretty rad stuff with blades that are largely stateless (more on this below.) We've extended upwards of 45% off (better at end of quarter) to a large portion of our customer base. Pretty much every customer that's bought them has fallen in love with them. I have one customer right now that buys ~10-11 chassis worth every 8 months or so. The CPU and RAM configs are based on the whole "virtual DIMM" extended memory tech that Cisco's pushed. It's not on every platform but they do have some configs you can't get any other way and in those configs you don't have to clock the memory down when you max the platforms out. I'd say the only legitimate con here is that Cisco's logistics are absolutely horrible. That said we've been able to get most orders out within a week of getting a PO. Some things we're doing with UCS: 1. I can automate the provisioning and identity of blades. This goes down to every last detail including how many NICs, what MACs they have, how many HBA's, what WWWNs and even what UUID the blades get stamped with. This is helpful for us for our provisioning systems since we store metadata in all the unique identifiers. If you want a Hyper-V server the MAC is X, if it's an ESXi server the MAC is Y. Based on this data we feed a specific PXE image to the box to stand it up. 2. I can store BIOS settings and firmware versions in a blade profile. This is helpful for large scale consistent deployments and/or changes. This is also helpful when you need to change a BIOS setting over a large number of blades since it can propagate that change down to all the blades using that policy. 3. I boot everything from SAN; what this means is I can replicate my server boot LUNs to the DR site and be 100% confident that it's going to boot at the DR site. I export my server profiles to my DR site and apply them to the UCS there. For all intents and purposes that blade looks exactly the same as the blade in my primary right down to the 'dmidecode' output. This is helpful for some of the workloads that haven't yet been virtualized. 4. Hardware upgrades take almost no time at all. I buy a new blade, pop it in a chassis then just move a server profile to it. If for whatever reason I have a problem I can roll it back pretty quickly by just moving the server profile back. 5. (this is more for the reseller side of our business); I can sell installation services for dirt cheap because we've completely automated the provisioning process (thanks to the UCSM API); that means we spend a day or so racking all of your chassis then about an hour doing the logical config on the blade. Then it's a day of walking the customer through how things work, etc. 6. It's a single management point for up to 320 blades (depending on oversub ratios.) Practically speaking we generally see on average ~12-16 chassis per pair of fabric interconnect though. 7. Unified port and 10 gig. Plug your NAS/iSCSI appliance right into the fabric interconnect if you're not yet on 10 gig. Optionally run the fabric interconnects in full fabric mode. 8. Having a read only NX-OS instance running on the FI is pretty helpful for troubleshooting. 9. Boot from software iSCSI! 10. I can actually manage my rackmount servers from the fabric interconnects that manage my blades. Backing up and restoring UCS configs is a lot more straightforward to me than it is on HP. I could go into more details with respect to the benefits if I knew more about your environment. I'm not saying you should absolutely buy UCS but it's good to evaluate it on it's merits (we have no trouble getting a 4 blade config with a pair of FIs in the customer doors to eval for 90 days.) Feature wise it's got most everything I can think of that HP does. Some actual cons: 1. Java.... I'm tired of dealing with java. Cisco is no better with their java apps and sometimes java updates break UCSM. 2. If you don't take time to plan things up front and understand how UCS behaves you can very easily paint yourself in a corner. 3. call home isn't configured by default. 4. if you're using brocade switches for northbound FC connectivity you may run into some issues with load distribution. Toss some questions out if you're interested in more information. 1000101 fucked around with this message at 10:25 on Nov 9, 2013 |
|
#
?
Nov 9, 2013 08:32
|
|
|
And the cabling. Oh my god the cabling. So clean. So fast. So easy.
|
|
#
?
Nov 9, 2013 16:03
|
|
|
Dr. Arbitrary posted:Anyone know much about UCS? It's a GUI monitoring system that allows you to check on the status of hosts and connections in your DC. I use it at work and monitor our racks of Cisco blades. I haven't had any real formal training with it, but I assume UCS is utilizing VMware's API to run a continuous poll on the health of each physical system, hard drives, memory, network links and processors. If there's a problem with anything you see a number under a color indicator - I'd have to fire up my work machine to tell you which colors indicate what specifically, but basically anything under red is something that needs attention.
|
|
#
?
Nov 9, 2013 20:10
|
|
|
If you want to get a feel for the UI you can access the platform emulator here: https://developer.cisco.com/web/unifiedcomputing/ucsemulatordownload Will give you a functional UI (you can't boot blades though) and an option to fiddle with the API.
|
|
#
?
Nov 9, 2013 20:50
|
|
|
Dilbert As gently caress posted:Wow actually starting to like citrix in some respect the more I deploy it vs other engineers. Citrix owns if it's deployed correctly and kept up to date (and I used to hate it and was probably one of the biggest Anti-Citrix guys out there, but it was because a lot of Citrix admins suck so I assumed Citrix sucked). And, honestly, XenApp <=6.5 (IMA) is pretty lame compared to XenDesktop 7 (FMA). If they were separate products by two neutral companies, View would be in a lot of trouble. View rides vSphere's coattails. Edit: I'm assuming by Citrix you mean XenApp/XenDesktop and not XenServer. If you mean XenServer then 
|
|
#
?
Nov 10, 2013 00:19
|
|
|
Citrix products generally have great features, but are difficult to deploy and admin properly. Like you said, three, most Citrix admins are terrible. I have a couple of XenApp/XenDesktop/XenServer environments that I am involved with, but because we are a small MSP they were set up wrong and time has never been taken to fix them. It's really frustrating because I had a decent sized Citrix environment that I implemented and administered before coming to work here, and it was great. I really miss Provisioning Services.
|
|
#
?
Nov 10, 2013 00:26
|
|
|
three posted:Citrix owns if it's deployed correctly and kept up to date (and I used to hate it and was probably one of the biggest Anti-Citrix guys out there, but it was because a lot of Citrix admins suck so I assumed Citrix sucked). And, honestly, XenApp <=6.5 (IMA) is pretty lame compared to XenDesktop 7 (FMA). quote:Edit: I'm assuming by Citrix you mean XenApp/XenDesktop and not XenServer. If you mean XenServer then Xenserver can kiss my rear end, it's like ESXi 3.5 but way more broken
|
|
#
?
Nov 10, 2013 05:17
|
|
|
I like you 1000101 but Imma need to nitpick you.1000101 posted:Cisco's cutting some pretty deep discounts for them (smartplay bundles get you in the door for dirt cheap). It's pretty price competitive against HP C7000 and featurewise I'd put it ahead. It's pretty much the farthest thing from a whitebox HP with "a custom java management console." UCSM is pretty damned extensible and we've done some pretty rad stuff with blades that are largely stateless (more on this below.) Did this change in the latest revision? the onboard poo poo still felt laggy as poo poo in G2/G1; I'd imagine they are on G4 as they follow intels click-clock cycle. I thought HP still made their boards which was one of the many reasons they introduced supported virtual CUCM on HP first, as a hand off. quote:We've extended upwards of 45% off (better at end of quarter) to a large portion of our customer base. Pretty much every customer that's bought them has fallen in love with them. I have one customer right now that buys ~10-11 chassis worth every 8 months or so. Their blades, I won't knock you, they are drat good and converged networking makes them sweet as butter. My complaint is, excluding Med-large Enterprise with constantly growing or somewhat large scale TI staff they still fall short of the SMB base and to some extent the Medium-small enterprise workspace. Granted they have some nice features but many other vendors such as HP/Dell/IBM can provide 70-85% the level of functionality that UCSM provides. I guess it really depends on the requirements of the customer to find if UCSM is viable to the customer needs. quote:The CPU and RAM configs are based on the whole "virtual DIMM" extended memory tech that Cisco's pushed. It's not on every platform but they do have some configs you can't get any other way and in those configs you don't have to clock the memory down when you max the platforms out. Depends on the "clock down" I'd argue clock down of ram vs CAS I can get on other boxes but I think I see where you are going with it but not completely sure. The whole "we don't do AMD" makes me a bit ehh. I mean yeah AMD lags behind on a few things but Piledriver did step it up quite a bit. quote:I'd say the only legitimate con here is that Cisco's logistics are absolutely horrible. That said we've been able to get most orders out within a week of getting a PO. I too love adding a embedded usb drive of 8GB for 125$ and having a shipment delayed for 21 days quote:1. I can automate the provisioning and identity of blades. This goes down to every last detail including how many NICs, what MACs they have, how many HBA's, what WWWNs and even what UUID the blades get stamped with. This is helpful for us for our provisioning systems since we store metadata in all the unique identifiers. If you want a Hyper-V server the MAC is X, if it's an ESXi server the MAC is Y. Based on this data we feed a specific PXE image to the box to stand it up. I agree with you on most of that, I can see use case in extra large environments with PCI/gove/healthcare/data sec/etc audits. quote:9. Boot from software iSCSI! Come on man autodeploy + failback to embedded USB; Or are you talking large+ scale gov/healthcare deploys if so Okay I can see SW iSCSI boots as a goody but autodeploy man... quote:Backing up and restoring UCS configs is a lot more straightforward to me than it is on HP. [quote] Agree'd I think a bunch of it comes down to customer needs and what value they find in X features
|
|
#
?
Nov 10, 2013 05:37
|
|
|
We use UCS where I work. I can say that for our most recent refresh, I recommended going with them, despite my being the most cost conscious member of my team. We did not pay an outrageous premium, in fact, I believe they were less than 10% more than the "equivalent" HP, and that was without the HPs being licensed for ILO. I would really prefer if Cisco offered an AMD option, since we found that AMD provided a better price/performance for VDI, but even being Intel only they are a pretty good value.
|
|
#
?
Nov 10, 2013 05:43
|
|
|
Dilbert As gently caress posted:I still like view for LAN, CAD, medical, or BYOD(pluginless HTML access to a desktop is loving cool as poo poo). But Citrix has some cool poo poo like program redirect and other things. Citrix has had an HTML5 receiver for awhile: http://blogs.citrix.com/2012/08/31/receiver-for-html5-is-now-available/ XenDesktop's vGPU has outperformed View's as well (requires XenServer, though): http://blogs.gartner.com/gunnar-berger/understanding-virtual-desktop-vdi-gpu-technologies/
|
|
#
?
Nov 10, 2013 05:54
|
|
|
Dilbert As gently caress posted:I like you 1000101 but Imma need to nitpick you. Haven't seen issues with it since UCS 1.4 or even previous to that. Customers using HP C7000 have compared the performance favorably. In my lab right now I don't notice any difference in lag/performance between the UI of OA vs UCSM. Regarding CUCM I would rather drink a vat of bleach before I ever touch anything involving voice. I'm guessing the certification process was due more to HP starting the process sooner. When UCS was conceived it was done so in a separate spin-out which was later spun in. quote:Their blades, I won't knock you, they are drat good and converged networking makes them sweet as butter. My complaint is, excluding Med-large Enterprise with constantly growing or somewhat large scale TI staff they still fall short of the SMB base and to some extent the Medium-small enterprise workspace. Granted they have some nice features but many other vendors such as HP/Dell/IBM can provide 70-85% the level of functionality that UCSM provides. I guess it really depends on the requirements of the customer to find if UCSM is viable to the customer needs. I'm not sure I follow how they "fall short of the SMB base and to some extend the medium-small enterprise workspace." Our VAR side has made some pretty good inroads into the SMB marketplace since, as I've said before, we can get pretty price competitive with Cisco's aggressive discounts. I'd argue as an SMB you're in a more favorable place to continue down the cisco road for growth since your largest investment is going to be the fabric interconnects. The UCS Chassis and IOMs are pretty cheap. That just leaves the blades which without aggressive discounts cost the same as HP counterparts. So now I've got a bunch of stuff to make my life easier at the same or less cost. Why not go with a more feature-rich solution? quote:Depends on the "clock down" I'd argue clock down of ram vs CAS I can get on other boxes but I think I see where you are going with it but not completely sure. The whole "we don't do AMD" makes me a bit ehh. I mean yeah AMD lags behind on a few things but Piledriver did step it up quite a bit. I haven't bothered with an AMD server since Intel launched Nehalem years ago. That said, sticking with 1 vendor does help Cisco keep costs down and I personally don't care what CPU is in there as long as it meets customer needs. If tomorrow a 1000 dollar AMD ran say 15% faster than a 1000 dollar Intel I might decide to care. Regarding RAM speeds I've run into applications where this matters. Large data sets people tend to want all of the memory bandwidth available to them. quote:I too love adding a embedded usb drive of 8GB for 125$ and having a shipment delayed for 21 days I don't think we've ever let an order get held up for an embedded USB drive. Who were you using for distribution? quote:I agree with you on most of that, I can see use case in extra large environments with PCI/gove/healthcare/data sec/etc audits. A lot of that we've developed for the SMB market where the VMware guy is also the server guy, windows guy and storage guy. Anything we can do to make the "lone ranger" sysadmin job easier is generally a good thing. Some of it (the DR stuff for example) was built for larger enterprise. quote:Come on man autodeploy + failback to embedded USB; Or are you talking large+ scale gov/healthcare deploys if so Okay I can see SW iSCSI boots as a goody but autodeploy man... I'm talking about things other than ESXi servers? Hyper-V? Citrix? KVM? Just running stuff on bare metal? This was a pretty useful feature in a software dev lab that was pretty deep into NetApp storage. quote:I'd like to argue, keep the HW platform as low config and standered as I can but I think I see where you are going on that. If that's the case then use rackmount systems and avoid blades entirely. If we're comparing apples to apples though where you have things like virtual connect/flex and you're carving out network cards and HBAs then I think people can appreciate the UCS backup process. quote:Agree'd I think a bunch of it comes down to customer needs and what value they find in X features If it does everything that competing platform does, and can cost less than the competing platform and also brings additional things to the table that the competing platform doesn't then why not look at it as a serious competitor?
|
|
#
?
Nov 10, 2013 06:03
|
|
|
Dilbert As gently caress posted:Their blades, I won't knock you, they are drat good and converged networking makes them sweet as butter. My complaint is, excluding Med-large Enterprise with constantly growing or somewhat large scale TI staff they still fall short of the SMB base and to some extent the Medium-small enterprise workspace. Granted they have some nice features but many other vendors such as HP/Dell/IBM can provide 70-85% the level of functionality that UCSM provides. I guess it really depends on the requirements of the customer to find if UCSM is viable to the customer needs. I'm nitpicking your nitpick at this point, but come on. Of course SMB does not need the complexity of UCS because managing a handful servers and a couple off-brand switches is not complicated. When you're talking about hundreds(++) of servers being able to manage all of that from within one app becomes a lot more attractive. Also, yes, gently caress Java forever.
|
|
#
?
Nov 10, 2013 06:07
|
|
|
Docjowles posted:I'm nitpicking your nitpick at this point, but come on. Of course SMB does not need the complexity of UCS because managing a handful servers and a couple off-brand switches is not complicated. When you're talking about hundreds(++) of servers being able to manage all of that from within one app becomes a lot more attractive. If there's one thing we can all agree on!
|
|
#
?
Nov 10, 2013 06:21
|
|
|
Docjowles posted:I'm nitpicking your nitpick at this point, but come on. Of course SMB does not need the complexity of UCS because managing a handful servers and a couple off-brand switches is not complicated. When you're talking about hundreds(++) of servers being able to manage all of that from within one app becomes a lot more attractive. The more I typed a repsonse to 1000101, yeah I agree. I really want to move out of this SMB hell hole I live in. I have and do scale large virtual environments but, a large part of my job still hold me back to the S of SMB; of which I really want to get out of and push what I know/don't know. Yes defense has a large hold on where I work at, no I do not want to be a military IT jocky for the rest of my career; and gently caress no I don't want to live in this area. Hope to see you at pex 1000101, I owe you dinner. three posted:Citrix has had an HTML5 receiver for awhile: http://blogs.citrix.com/2012/08/31/receiver-for-html5-is-now-available/ I was under the impression that the HTML5 reciever still required a client side plugin install? Is that not the case anymore? Also yeah but I don't know anyone deploying or recommending deploying Xenserver Dilbert As FUCK fucked around with this message at 07:03 on Nov 10, 2013 |
|
#
?
Nov 10, 2013 06:59
|
|
|
1000101 posted:I don't think we've ever let an order get held up for an embedded USB drive. Who were you using for distribution? I blame commerce workspace for this, as soon as you add a part which needs to be installed before the product ships (and isn't a required item) it pushes the estimated ship date out 3 weeks.
|
|
#
?
Nov 10, 2013 13:59
|
|
|
Posted this a few weeks ago but I think it's worth a mention if you're diving in to UCS right now: Java 7 Update 45 breaks the Cisco UCS console: http://tools.cisco.com/Support/BugToolKit/search/getBugDetails.do?method=fetchBugDetails&bugId=CSCuj84421 Workaround according to Cisco? Use Java version 7 update 40 or below. Earlier releases are available from: http://www.oracle.com/technetwork/java/javase/downloads/java-archive-downloads-javase7-521261.html Anyway, I think the blade profile capabilities of Cisco are pretty drat nice. Configure your BIOS and firmware settings, from a Java webpage, click apply and trigger a reboot and boom, whole new BIOS, NIC firmware, or BIOS settings, or anything. Lose a blade? Pop a hot spare in that slot and it will come up with the EXACT same settings / environment as the dead one. The OS or VMware will never even know the difference. You've got pools for MAC address, management IP, WWN, everything, so deploying additional blades is really easy. They've got an API for some seriously heavy scripting capabilities and the command line controls are pretty close to what you get on a Nexus or MDS switch so your network guys will like it. I will say the KVM and ISO mounting sucks compared to Dell DRAC 6 / 7 though. And the out of band RAID management capabilities blow for the Cisco C series rackmount servers. In fact, I couldn't get them to tell me if a RAID rebuilt properly or not, I'd have to reboot and check the BIOS. As always, I'm sure all the vendors have big warchests to go out and win deals from competitors. If you press them all hard with pricing from HP, Dell, and Cisco, you can get a pretty good deal. I'd be shocked if the total price band between all three was more than 20%. hackedaccount posted:Nah because organizations that need true fault tolerance would drop the cash on something like these (and they support vSphere too). Dilbert As gently caress posted:Don't worry guys vSMP FT is coming soon******* Our company ripped our Stratus FT boxes out, too many problems with them. We decided we could live with a 30 second VM reboot. Anyone want them? I can let them go for cheap. I've been trying to find a decent reseller to get rid of them. You know, I've been thinking about FT lately. Mellanox has a $700 twin 40 GbE / 56Gb Infiniband NIC for like $700. If they make crossover cables, you could pipe 80-112Gb between two boxes which would sure as hell fix any bandwidth problems to FT with multiple threads. They've got switches starting at $16K for ~30 ports, but no SFPs and I bet they're going to charge you for port licenses and all kinds of other hidden fees / prices since that's just too cheap to believe. I have a feeling Infiniband is eventually going to be a pretty popular interconnect on top of rack switches, especially with storage getting pushed to the "edges" of your network. Sooner or later, we'll all have like 16U of storage and 24U of compute, and 2U of networking, per rack. No more monolithic racks of all storage or all compute. El_Matarife fucked around with this message at 17:56 on Nov 10, 2013 |
|
#
?
Nov 10, 2013 17:46
|
|
|
I may already have asked this, but how well does USB passthrough work on an ESXi host? If I had a USB device that flashed microcontrollers, could I hook that up to an ESXi host and give my development VM access to it? I feel like this should be easy enough to test, but I don't have access to my ESXi server at work right now.
|
|
#
?
Nov 10, 2013 20:04
|
|
|
Martytoof posted:I may already have asked this, but how well does USB passthrough work on an ESXi host? If I had a USB device that flashed microcontrollers, could I hook that up to an ESXi host and give my development VM access to it? I've only used it for a 13 year old dongle and that worked fine.
|
|
#
?
Nov 10, 2013 23:07
|
|
|
I'm starting to look at build some sort of home server so that I can play around with ESXi and Hyper-V at home, and am trying to figure out what the best approach would be. What are the benefits of purchasing a specific "server" model from Newegg vs basically just building a PC and jamming 32 gigs of RAM into it? Would I be bottlenecking myself somehow with the PC approach? Obviously fancy high availability options such as redundant NIC's and such aren't necessary for this purpose. Just looking for some general guidance and wanting to make sure I'm not going down the wrong path.
|
|
#
?
Nov 11, 2013 02:28
|
|
|
Martytoof posted:I may already have asked this, but how well does USB passthrough work on an ESXi host? If I had a USB device that flashed microcontrollers, could I hook that up to an ESXi host and give my development VM access to it? Check out the digi USB Anywhere devices. That's what I use to provide USB access to VMs.
|
|
#
?
Nov 11, 2013 02:39
|
|
|
parid posted:Check out the digi USB Anywhere devices. That's what I use to provide USB access to VMs. That's not a terrible idea (and honestly probably a cleaner idea in the long term), but I'm not going to spend $200-something if ESXi can pass through USB well enough to do the job 
|
|
#
?
Nov 11, 2013 03:45
|
|
|
CtrlMagicDel posted:I'm starting to look at build some sort of home server so that I can play around with ESXi and Hyper-V at home, and am trying to figure out what the best approach would be. What are the benefits of purchasing a specific "server" model from Newegg vs basically just building a PC and jamming 32 gigs of RAM into it? Would I be bottlenecking myself somehow with the PC approach? Obviously fancy high availability options such as redundant NIC's and such aren't necessary for this purpose. Just looking for some general guidance and wanting to make sure I'm not going down the wrong path. Whitebox build is interesting to say the least. I went ahead and built my own whitebox (i7-4770, 32GB RAM, Raid Controller + 2x300GB WD Blues in a RAID 1, 2x2TB Seagate NAS drives in a RAID 0 and set up as RDM dedicated to my Linux VM, and 2 NIC cards, giving me a total of 3 NICs). It's not the flashiest ESXi box on the block, but considering I only run a firewall, Linux box and Windows box, it is plenty powerful for what I need. In the next few days I'm going to be building out a nested ESXi setup for labbing for my VCP5 exam, but it should still be able to handle everything and still have resources left over (although it might get a little tight). I've actually been having quite a bit of fun with it, and of course work provides me a ton of opportunity to get additional practical experience in a huge virtual environment.
|
|
#
?
Nov 11, 2013 03:57
|
|
|
CtrlMagicDel posted:I'm starting to look at build some sort of home server so that I can play around with ESXi and Hyper-V at home, and am trying to figure out what the best approach would be. What are the benefits of purchasing a specific "server" model from Newegg vs basically just building a PC and jamming 32 gigs of RAM into it? Would I be bottlenecking myself somehow with the PC approach? Obviously fancy high availability options such as redundant NIC's and such aren't necessary for this purpose. Just looking for some general guidance and wanting to make sure I'm not going down the wrong path. Basically, ESXi in particular has really iffy support for consumer chipsets (especially newfangled stuff, where some USB3.0 chipsets still hang the boot process for 10 minutes then never work) and you're basically required to buy Intel NICs. Hyper-V takes (more or less) Windows driver support, so you're safer there (along with Xen[Server], KVM, et al), but ESXi whiteboxes are "server hardware, whitebox HCL wiki, or gamble on support".
|
|
#
?
Nov 11, 2013 04:30
|
|
|
I've never had an issue with a whitebox build running esxi. Just have an intel PCI express NIC on hand. I don't see the value in spending extra for a "server" for a home build anyway. Once it's running esxi, you are not going to mess with it outside of upgrades anyway.
|
|
#
?
Nov 11, 2013 04:59
|
|
|
Martytoof posted:That's not a terrible idea (and honestly probably a cleaner idea in the long term), but I'm not going to spend $200-something if ESXi can pass through USB well enough to do the job Are you just trying to plug in a device that is local to the ESX box? If so, that should be just fine. Nothing special is needed. I have done that with wireless network adapters connecting to a Windows VM.
|
|
#
?
Nov 11, 2013 05:39
|
|
|
Daylen Drazzi posted:Whitebox build is interesting to say the least. I went ahead and built my own whitebox (i7-4770, 32GB RAM, Raid Controller + 2x300GB WD Blues in a RAID 1, 2x2TB Seagate NAS drives in a RAID 0 and set up as RDM dedicated to my Linux VM, and 2 NIC cards, giving me a total of 3 NICs). It's not the flashiest ESXi box on the block, but considering I only run a firewall, Linux box and Windows box, it is plenty powerful for what I need. In the next few days I'm going to be building out a nested ESXi setup for labbing for my VCP5 exam, but it should still be able to handle everything and still have resources left over (although it might get a little tight). Something like that seems more than good enough. Since I'm not going to be doing anything that complicated right away, I'm guessing I could skimp on the RAID controller and additional NIC's at least initially, right? Otherwise, any suggestions for a cheap RAID controller? evol262 posted:Basically, ESXi in particular has really iffy support for consumer chipsets (especially newfangled stuff, where some USB3.0 chipsets still hang the boot process for 10 minutes then never work) and you're basically required to buy Intel NICs. Hyper-V takes (more or less) Windows driver support, so you're safer there (along with Xen[Server], KVM, et al), but ESXi whiteboxes are "server hardware, whitebox HCL wiki, or gamble on support". Yeah, I've heard that in various places which is disappointing given my work is all ESXi and I'm looking for some experience installing it from scratch. Though getting some Hyper-V experience might be nice as well given I've suggested that it might be worth doing a cost savings analysis down the road for potentially moving less critical servers to Hyper-V, or at least bringing it into our lab environment to compare it to ESXi. adorai posted:I've never had an issue with a whitebox build running esxi. Just have an intel PCI express NIC on hand. I don't see the value in spending extra for a "server" for a home build anyway. Once it's running esxi, you are not going to mess with it outside of upgrades anyway. Cool, that seems like a cheap and easy enough solution.
|
|
#
?
Nov 11, 2013 05:39
|
|
|
I only went with the RAID controller because I wanted to make drat sure that my datastore had redundancy in the event of a hard drive failure. As for the RAID 0 hooked to my Linux VM, my intention is to buy a 4TB Seagate NAS drive and back those drives to it. The Linux box is a pure file storage system for all my movies and music, so I'd like to make sure I don't lose anything, but even if I did it's not like it would be the end of the world. I'm also leaning towards buying a NAS appliance, maybe one of the 4-bay Netgear units and setting it up as my dedicated storage, but I don't happen to have several hundred bucks to drop on it right now, so that's something for next year. I find that the more I get into virtualization the more I find I want to learn about and try out on my home system. If only I could hit the lottery...
|
|
#
?
Nov 11, 2013 13:49
|
|
|
DevNull posted:Are you just trying to plug in a device that is local to the ESX box? If so, that should be just fine. Nothing special is needed. I have done that with wireless network adapters connecting to a Windows VM. Yeah. It's not ideal since I really shouldn't be going into our server closet to do stuff like flashing microcontrollers but I've got my development VM on my testbed ESXi server and I'm just too lazy to reinstall everything on a local VM. All I'll be doing is plugging in an AVR Dragon or Olimex ARM JTAG adapter into the back of the ESXi host itself and passing them through to a VM to use in their respective IDEs. I've just never actually passed a USB device through before so I wanted to make sure it was possible or whether I was mistaken and/or making poo poo up. Thanks! 
|
|
#
?
Nov 11, 2013 14:19
|
|
|
adorai posted:I've never had an issue with a whitebox build running esxi. Just have an intel PCI express NIC on hand. I don't see the value in spending extra for a "server" for a home build anyway. Once it's running esxi, you are not going to mess with it outside of upgrades anyway. I'm not saying it's worth it to spend extra for a "server" for a home lab, but it probably is worth it to check the whitebox HCL, because boot "hang" issues aren't exactly uncommon. CtrlMagicDel posted:Yeah, I've heard that in various places which is disappointing given my work is all ESXi and I'm looking for some experience installing it from scratch. Though getting some Hyper-V experience might be nice as well given I've suggested that it might be worth doing a cost savings analysis down the road for potentially moving less critical servers to Hyper-V, or at least bringing it into our lab environment to compare it to ESXi. I've never understood the "installing it from scratch" argument for anything. A secretary could do it.
|
|
#
?
Nov 11, 2013 15:35
|
|
|
evol262 posted:I'm not saying it's worth it to spend extra for a "server" for a home lab, but it probably is worth it to check the whitebox HCL, because boot "hang" issues aren't exactly uncommon. Good advice, I'll try to find a compatible motherboard right off the bat. I suppose just the installation process itself isn't particularly valuable, but everything else that comes with it. I also feel more comfortable with a technology when I've walked through the installation/configuration of it myself, as minimal as that might be.
|
|
#
?
Nov 11, 2013 17:54
|
|
|
So I'm a little confused about labs for the Stanly class. I just whipped through 10 of them in a 6 hours of reservation. I was worried that I'd have to be jockeying for time to get all the labs done, but if I look for the next few days there are zero reservations. Aren't we all sharing the same two pods in the class? Is everybody just lazy and not doing anything with the class yet? I assume that we have two pods at least, I can only see ICM_POD_241 and ICM_EQ3_POD15. Does it matter which I choose? Also how do they know we've completed the labs?
|
|
#
?
Nov 11, 2013 22:57
|
|
|
FISHMANPET posted:So I'm a little confused about labs for the Stanly class. I just whipped through 10 of them in a 6 hours of reservation. I was worried that I'd have to be jockeying for time to get all the labs done, but if I look for the next few days there are zero reservations. Aren't we all sharing the same two pods in the class? Is everybody just lazy and not doing anything with the class yet? I assume that we have two pods at least, I can only see ICM_POD_241 and ICM_EQ3_POD15. Does it matter which I choose? a) I never once saw another reservation in all my time scheduling labs b) I think it's mostly the honour system. I think the system only tracks the labs you've scheduled so make sure you let the instructor know you're doing multiple labs via the messaging system so he's not like "yo you only did 4 labs buddy" when you've actually finished them all. This class is basically a rubber stamp, which isn't too bad in and of itself. I really kind of expected that and wasn't too blown away when it felt so easy and ghetto.
|
|
#
?
Nov 12, 2013 02:09
|
|
|
Think I might look up a bit more citrix VDI stuff, it's pretty decent if set up right. Any good resources? Only ones I found on amazon that aren't 3 years old are http://www.amazon.com/Getting-Started-Citrix-VDI---Box/dp/1782171045/ref=sr_1_2?ie=UTF8&qid=1384218625&sr=8-2&keywords=citrix+vdi there is also http://www.amazon.com/Citrix-XenDes...ix+xendesktop+7 but it seems a bit dated I mean I understand most the concepts already but just want to finite it. I'd like to get a bit more proficient at it before I move.
|
|
#
?
Nov 12, 2013 02:14
|
|
|
Martytoof posted:a) I never once saw another reservation in all my time scheduling labs I thought I saw a message in the FAQ saying that "yes, we can see everything you've done in the lab, so there's no need to contact the instructor unless you have a question." Basically I've finished up lab 13, so I'm just waiting for the instructor to open the last three modules so I can knock them out as well, and then I should be done.
|
|
#
?
Nov 12, 2013 03:08
|
|
|
FISHMANPET posted:So I'm a little confused about labs for the Stanly class. I just whipped through 10 of them in a 6 hours of reservation. I was worried that I'd have to be jockeying for time to get all the labs done, but if I look for the next few days there are zero reservations. Aren't we all sharing the same two pods in the class? Is everybody just lazy and not doing anything with the class yet? I assume that we have two pods at least, I can only see ICM_POD_241 and ICM_EQ3_POD15. Does it matter which I choose? I am pretty sure they never check if you completed the labs. I took it last session and emailed the dude to confirm that the only things they grade on are the labs and the quizzes the teacher said yes and then offered to reset my environment. I hadn't even finished half way. I got the impression that they probably just mark everyone who completes quizzes as completing and make you eligible to take the test. The reservations are just to ensure that you don't leave it running when not using I think. I have never seen another reservation listed.
|
|
#
?
Nov 12, 2013 03:09
|
|
|
|
| # ? May 8, 2024 06:27 |
|
|
demonachizer posted:I am pretty sure they never check if you completed the labs. I took it last session and emailed the dude to confirm that the only things they grade on are the labs and the quizzes the teacher said yes and then offered to reset my environment. I hadn't even finished half way. I got the impression that they probably just mark everyone who completes quizzes as completing and make you eligible to take the test. Quoted from the Class Sections part of the course website: quote:Q. Is lab progress tracked?
|
|
#
?
Nov 12, 2013 03:42
|
|