|
thebigcow posted:http://www.youtube.com/watch?v=YwASii2f5c8 Come on, man. I had a rough night, and I hate the fuckin' Eagles, man!
|
#
?
May 11, 2013 06:00
|
|
|
|
| # ? May 11, 2024 19:44 |
|
|
Docjowles posted:Come on, man. I had a rough night, and I hate the fuckin' Eagles, man! *kicks you out onto the street* (although I will have that White Russian)
|
|
#
?
May 12, 2013 15:38
|
|
|
What VMware products do you guys use? My company has been pitching around the idea of putting a lab for engineers to study in, and PoC/demo up for customers. They liked my super micro fat twin "all-in-one" pitch quite a lot. I was thinking about demo'ing vCenter(duh) Enterprise Plus features View suite SRM Veeam/VDP vCoPs vCloud Director(debating) Any other products out there you guys use?
|
|
#
?
May 16, 2013 16:29
|
|
|
Did the vcenter 5.0-->5.1 upgrade using the simple installation i think i've made a horrible mistake edit: its a small environment (4 hosts) and I'm seriously considering just whacking the entire vcenter installation one weekend and installing the appliance instead. Alctel fucked around with this message at 23:23 on May 17, 2013 |
|
#
?
May 17, 2013 23:19
|
|
|
Alctel posted:Did the vcenter 5.0-->5.1 upgrade Simple install is a great way to gently caress poo poo up, I don't think VMware recommends doing it anymore outside of PoC. I wonder if the VA for vCenter 5.1 u1 supports MSSQL, I know vCloud director now has the option.
|
|
#
?
May 18, 2013 00:56
|
|
|
Alctel posted:Did the vcenter 5.0-->5.1 upgrade Did this at my old place not long after 5.1 came out. Took about a day to get things happy but was smooth sailing after. Little more nervous at my new place since other people know what vCenter is and use it.
|
|
#
?
May 18, 2013 03:17
|
|
|
Moey posted:Did this at my old place not long after 5.1 came out. Took about a day to get things happy but was smooth sailing after. Knowing what something is and knowing how to administer it are two completely separate things. Just because someone "knows" what X is doesn't mean they understand it.
|
|
#
?
May 18, 2013 03:22
|
|
|
I think he means that if there was a failed upgrade and a couple days downtime, people would notice and complain. Get a good fall back plan together. It could only take an hour to get it all back up after a failed install if all you have is a db backup. You could also "practice" the upgrade with copies of your center.
|
|
#
?
May 18, 2013 04:51
|
|
|
parid posted:I think he means that if there was a failed upgrade and a couple days downtime, people would notice and complain. Get a good fall back plan together. It could only take an hour to get it all back up after a failed install if all you have is a db backup. How does a failed upgrade result in "a couple of days of downtime"... I mean seriously, HOW does that happen? Does one not test things in a lab, or understand how upgrades work? I mean I realize poo poo can go bad in the worst poo poo but how do you not have a backup plan or know a 4-8hr fix something is wrong
|
|
#
?
May 18, 2013 04:54
|
|
|
Corvettefisher posted:How does a failed upgrade result in "a couple of days of downtime"... I mean seriously, HOW does that happen?
|
|
#
?
May 18, 2013 04:55
|
|
|
Also, if for some reason you left your fate to VMware support. If you worked on their schedule without pushing hard, you would be down for days.
|
|
#
?
May 18, 2013 05:48
|
|
|
parid posted:Also, if for some reason you left your fate to VMware support. If you worked on their schedule without pushing hard, you would be down for days. Like what? Also does this mean you went #yolo on an update without reading on it or testing. If so I can tell you the software ain't to blame Wait... parid posted:Did my 5.1u1 updates today. Two vcenters and 40 or so hosts. No issues! First time in over a year that a vcenter patch hasn't forced me to do a full roll back at least once. I did get a bit of a panic on when the center update hung on "waiting for services to start" for over half an hour. It did eventually start. This is you... You mean to tell me you did not stage an update in a test, for your environment of this size? AND you have a TAM? protip! THIS IS A GREAT TIME TO USE THIS!!! Dilbert As FUCK fucked around with this message at 06:09 on May 18, 2013 |
|
#
?
May 18, 2013 05:53
|
|
|
So Yeah I hate power-CLI, I am debating writing my own program(s) to do things that should be done in vSphere, such as image building and Storage plugins as well as some others. I find it hard to believe that VMware can't make a product to do this via GUI. *note: I have no problem with Power-CLI other than the fact I think VMware should stop guessing why people don't use X feature but only make it available to power CLI users. All I know is Python and Java so I guess I could write it in one of those Dilbert As FUCK fucked around with this message at 01:02 on May 19, 2013 |
|
#
?
May 19, 2013 01:00
|
|
|
Corvettefisher posted:All I know is Python and Java so I guess I could write it in one of those
|
|
#
?
May 19, 2013 01:35
|
|
|
Misogynist posted:If PowerCLI is too much trouble for you, you're going to have a lot of fun trying to work with the SOAP APIs. It's no trouble at all for me, I just hate it all being bound to a CLI based setup. Object oriented syntax is fine, I just think it is dumb people ask "why isn't X used" when the only way to do it is power cli. I mean I know a bit of perl but I wouldn't say I am fluent in it. I also think VMware should be a bit more innovative on some fronts. I mean seriosuly want to do something more advance expect a "well power cli is for that!". Call me spergy but that is a terrible excuse. Dilbert As FUCK fucked around with this message at 01:59 on May 19, 2013 |
|
#
?
May 19, 2013 01:46
|
|
|
Corvettefisher posted:This is you... You mean to tell me you did not stage an update in a test, for your environment of this size? AND you have a TAM? protip! THIS IS A GREAT TIME TO USE THIS!!! I have absolutely no idea what you are talking about. One of the vcenters was a test environment? You know nothing about my environment. Why so aggressive? I was trying to find other people with the TAM service as we aren't seeing any value in ours. For example, there's nothing he can do to help us with upgrades like this. I was hoping some shared experience from someone else who has a successful relationship might help me find what's going wrong in ours.
|
|
#
?
May 19, 2013 03:15
|
|
|
parid posted:I have absolutely no idea what you are talking about. One of the vcenters was a test environment? You know nothing about my environment. Why so aggressive? Most of my customer base over the last 5 years have had a TAM. They are by and large the most worthless people I've ever had to work with at VMware. I've known exactly 1 TAM who actually was helpful out of a dozen or so. It seems to be a position that VMware stuffs "those who can't" or at least the very lazy.
|
|
#
?
May 19, 2013 06:39
|
|
|
Misogynist posted:If PowerCLI is too much trouble for you, you're going to have a lot of fun trying to work with the SOAP APIs. There's a ruby wrapper (rbvmomi) if that's your thing, it's used in the Chef vSphere plugins.
|
|
#
?
May 19, 2013 10:20
|
|
|
We ran into something interesting recently. Well interesting to me, I am pretty much a vmware newbie (might actually get some training on this soon though...). We have two VMs running oracle databases. Test and Prod basically. Both have 8 cores assigned, however prod was 1 socket x 8 core, the test box 2 sockets x 4 core. The prod box started running slow 'around' the same time we did an update from 5.0 to 5.0 U2 a couple weeks back. A job that used to take hours now took days. Test didn't seem to be impacted. One of the server engineers saw that the test box had this slightly different config, and in a 'let me go home on a sunday afternoon' moment changed the prod box config to match test.. which fixed the performance issue. The time taken to run a particular job went from days to hours. Now I had been doing some reading and from what I understood, the socket vs cores configuration is allow getting around licensing for some OSes etc and is not meant to impact performance. However looking at the CPU ready stat, it dropped from ~5% (or higher) to ~1-2% after making that change, and the job itself is running orders of magnitude faster. Any thoughts on what can cause this? The confusing part for me is that under the original version of vmware we did not have any performance issues reported. The slow performance appeared to be related to the upgrade we did, but not 100%. Perhaps there was some change to the cpu scheduler between those versions?
|
|
#
?
May 20, 2013 19:30
|
|
|
luminalflux posted:There's a ruby wrapper (rbvmomi) if that's your thing, it's used in the Chef vSphere plugins.
|
|
#
?
May 20, 2013 19:39
|
|
|
I've built up 4 test VM's for a new project on a subdomain of our parent domain. I am JUST now finding out about this stupid child-domain auth bug (which prevents authentication for accounts if you are in a parent domain) in the new VMware 5.1 Update 1 software. Every fix that I've seen is for vCenter, and not standalone vSphere (currently running free license). Does anyone know if VMware has provided a resolution for that product?
|
|
#
?
May 20, 2013 19:48
|
|
|
GanjamonII posted:We ran into something interesting recently. Well interesting to me, I am pretty much a vmware newbie (might actually get some training on this soon though...). We have two VMs running oracle databases. Test and Prod basically. Both have 8 cores assigned, however prod was 1 socket x 8 core, the test box 2 sockets x 4 core. I'm still a bit confused you refer to "making that change" but what exactly did you change? Wicaeed posted:I've built up 4 test VM's for a new project on a subdomain of our parent domain. I am JUST now finding out about this stupid child-domain auth bug (which prevents authentication for accounts if you are in a parent domain) in the new VMware 5.1 Update 1 software. You can create local accounts on the host, but other than this kb article I haven't heard of much for an AD fix. However, I thought that was only an issue for people using vCenter Dilbert As FUCK fucked around with this message at 20:42 on May 20, 2013 |
|
#
?
May 20, 2013 20:38
|
|
|
Well, this bullshit is apparently related to our inept Network Administrators who take 3 days to create a firewall rule that doesn't work  I log in as root and still get that error. edit: I also blame VMware as they seem to be unable to create a straight forward document that simply lists the required ports for JUST the management access from the vSphere client to an ESXi server (NO I'M NOT RUNNING VCENTER PLEASE DONT TELL ME THE PORTS I NEED FOR VCENTER)
|
|
#
?
May 20, 2013 20:57
|
|
|
GanjamonII posted:We ran into something interesting recently. Well interesting to me, I am pretty much a vmware newbie (might actually get some training on this soon though...). We have two VMs running oracle databases. Test and Prod basically. Both have 8 cores assigned, however prod was 1 socket x 8 core, the test box 2 sockets x 4 core. Maybe there's some undocumented change in that update to how VMs are split across NUMA nodes depending on how their vCPUs are configured? Or, it could be something as silly as reconfiguring the production machine required a reboot and that fixed something  Either way, 8 cores is a lot. How many cores do the physical CPUs have in the host?Corvettefisher posted:I'm still a bit confused you refer to "making that change" but what exactly did you change?
|
|
#
?
May 20, 2013 21:05
|
|
|
http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=1012382#ESXi_5.x This is complicated? It even has the whitepaper at the top of the page... Erwin posted:Maybe there's some undocumented change in that update to how VMs are split across NUMA nodes depending on how their vCPUs are configured? Or, it could be something as silly as reconfiguring the production machine required a reboot and that fixed something Oh derp. I see I read that as 2 hosts 1x8 core and another 2x4 cores. I've always wondered if there was some small changes made with selecting sockets and cores, but everyone I talk to says it is purely a licensing thing. That would be good to do in my lab tonight. Dilbert As FUCK fucked around with this message at 21:11 on May 20, 2013 |
|
#
?
May 20, 2013 21:06
|
|
|
Erwin posted:Maybe there's some undocumented change in that update to how VMs are split across NUMA nodes depending on how their vCPUs are configured? Or, it could be something as silly as reconfiguring the production machine required a reboot and that fixed something 8 cores is overkill for this application but I don't think anyone involved in the project really understood how to allocate resources for VMs when this environment was stood up. One of the recommendations from what I'm doing now will be to cut that down, but things move slowly here and the apps guys are kind of afraid of rocking the boat in the middle of the project that they are working on The physical hosts have 2x E7-2850 10 core processors (shows up as 40 logical processors within vsphere due to hyperthreading). There is ~50 VMs spread over seven hosts of this spec, and most of them are over provisioned in the same way. Some of the CPU ready times are pretty high looking to me (20+ %). The actual CPU utilization on the hosts is very low (10-20% tops). Edit - it could have been due to requiring a reboot, I need to check with that team if they had tried that earlier. GanjamonII fucked around with this message at 21:25 on May 20, 2013 |
|
#
?
May 20, 2013 21:22
|
|
|
GanjamonII posted:
You mean spiked or sustained? This is a pretty large issue, CPU ready does not always correlate to cpu usage. Whoever set it up did not properly assign vCPU's. Normally CPU ready is 1-5%, 10% is where you will start seeing some noticeable performance degration. Dilbert As FUCK fucked around with this message at 21:43 on May 20, 2013 |
|
#
?
May 20, 2013 21:30
|
|
|
GanjamonII posted:The physical hosts have 2x E7-2850 10 core processors (shows up as 40 logical processors within vsphere due to hyperthreading). There is ~50 VMs spread over seven hosts of this spec, and most of them are over provisioned in the same way. Some of the CPU ready times are pretty high looking to me (20+ %). The actual CPU utilization on the hosts is very low (10-20% tops). Jeez, yeah, that's pretty bad. The good news is that with 10 core processors, 8-core VMs aren't wide so it should be sitting in one NUMA node no matter what. The correct fix is obviously to resize all the VMs, but as a crutch you could use a combination of host affinity and NUMA node affinity to keep the DB VM on a physical CPU where the only other VMs are 1- and 2-core VMs to cut down on co-scheduling issues. That would be a nightmare to manage, though.
|
|
#
?
May 20, 2013 22:13
|
|
|
Corvettefisher posted:You mean spiked or sustained? This is a pretty large issue, CPU ready does not always correlate to cpu usage. Whoever set it up did not properly assign vCPU's. Normally CPU ready is 1-5%, 10% is where you will start seeing some noticeable performance degration. I rechecked and right now some of the VMs are sustained at 15% spiking higher than that, but I have seem 20+ in the past. Hopefully we can get the re-sizing of these things done sooner rather than later.
|
|
#
?
May 20, 2013 22:47
|
|
|
GanjamonII posted:I rechecked and right now some of the VMs are sustained at 15% spiking higher than that, but I have seem 20+ in the past.  Yeah don't give VM's more than they need if you can help it. You would be surprised how many servers run happy on 1-2 vCPU's.
|
|
#
?
May 20, 2013 22:54
|
|
|
Corvettefisher posted:You mean spiked or sustained? This is a pretty large issue, CPU ready does not always correlate to cpu usage. Whoever set it up did not properly assign vCPU's. Normally CPU ready is 1-5%, 10% is where you will start seeing some noticeable performance degration.
|
|
#
?
May 21, 2013 01:11
|
|
|
I'd say the reboot is more likely the cause of the fix than changing sockets to cores or vice versa. Reboot could have cleared up issues within the OS, or placed the VM onto a new host with less CPU issues.
|
|
#
?
May 21, 2013 01:24
|
|
|
Corvettefisher posted:
|
|
#
?
May 21, 2013 02:39
|
|
|
three posted:I'd say the reboot is more likely the cause of the fix than changing sockets to cores or vice versa.
|
|
#
?
May 21, 2013 03:37
|
|
|
This might be a dumb question but is there anything specific I need to do to let an NMS query VCenter Server? anything related to VCenter hardening? I can poll ESXi in a lab easy... vCenter? Not so much. I admit I haven't spent a ton of time on this I've been busy wiht higher pri stuff but any help would be appreciated. Thanks. Thanks.
|
|
#
?
May 21, 2013 04:27
|
|
|
At the risk of asking a really retarded question, is the Windows Firewall on the vCenter server allowing whatever ports your NMS server polls through?
|
|
#
?
May 21, 2013 05:26
|
|
|
Have any of you had an issue with the console on vSphere Client taking forever to refresh and open? Whenever I try to view the console session of my vCenter Server installation, it takes a really long time (sometimes 10 minutes) before it finally displays. I think that I set enough resources except I am unsure about the total video memory and whether that could be affecting it. What would be a good default there for a vCenter Install? EDIT: If I right click, open console, it sometimes comes up much faster but if I wait for the built in, it is terrible. Demonachizer fucked around with this message at 15:08 on May 21, 2013 |
|
#
?
May 21, 2013 14:50
|
|
|
demonachizer posted:Have any of you had an issue with the console on vSphere Client taking forever to refresh and open? Whenever I try to view the console session of my vCenter Server installation, it takes a really long time (sometimes 10 minutes) before it finally displays. I think that I set enough resources except I am unsure about the total video memory and whether that could be affecting it. What would be a good default there for a vCenter Install? Known bug if you're on 5.1 with no patches. See http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=2037408
|
|
#
?
May 21, 2013 17:54
|
|
|
Going back to my situation I had a couple of questions for y'all smart people in here 1) Are the thresholds of eg 10% for CPU ready you're mentioning applicable to larger 8 way VMs? I've seen some blogs online which seem to point to higher CPU ready being acceptable for larger VMs. Eg 40% for 8 core VM. The highest VM I can see right now is sustained at 35% but doesn't seem to be doing any actual work (1-3% CPU utilization). 2) The CPUs in my hosts are 2x 10 core, but with hyper threading shows as 40 logical cores in vsphere. For purposes of calculating over-commitment should I be using 20 or 40 for the physical cores? Eg I have 40 vcpus allocated on one of my hosts, is this 1:1 or 2:1 virtual:physical?
|
|
#
?
May 21, 2013 17:55
|
|
|
|
| # ? May 11, 2024 19:44 |
|
|
So VMware is a service provider now. http://www.vmware.com/products/datacenter-virtualization/vcloud-hybrid-service
|
|
#
?
May 21, 2013 18:16
|
|
