JarryG's Posts

There is no power-source redundancy, so if your server has 2 ps-units, one of them is probably not working as expected. You should investigate it and change faulty ps-unit if necessary. If you... See more...
There is no power-source redundancy, so if your server has 2 ps-units, one of them is probably not working as expected. You should investigate it and change faulty ps-unit if necessary. If your server comes with 1 ps-unit, then you can forget this message, as it is telling you nothing new (you have 1 ps-unit, so of course there can not be any ps-redundancy)...
Unfortunatelly, ESXi does *not* support soft-/bios-/fake-raid (which is exactly what you have on your motherboard). Only real hardware raid-controllers (and not all of them). I recommend you to c... See more...
Unfortunatelly, ESXi does *not* support soft-/bios-/fake-raid (which is exactly what you have on your motherboard). Only real hardware raid-controllers (and not all of them). I recommend you to check "Hardware compabitility list"...
Just a side-note: maybe for your type of application different product would be more suitable, i.e. VMware Workstation. It has quite good 3d/gpu support. I was able to run some quite intensive 3d... See more...
Just a side-note: maybe for your type of application different product would be more suitable, i.e. VMware Workstation. It has quite good 3d/gpu support. I was able to run some quite intensive 3d-games, 3DMark benchmark, play hd-video or flash in VM...
I would try to boot *only* with new memory modules, and in the very same slots where you had 16GB up to now. If server does not boot, something wrong is with these new memory-modules. If it boots... See more...
I would try to boot *only* with new memory modules, and in the very same slots where you had 16GB up to now. If server does not boot, something wrong is with these new memory-modules. If it boots normally, I'd try to move these new dimms into the other two memory slots (which were empty until now) and again boot up with only these new modules. BTW, are all (both old and new) dimms of the very same type/model?
Hi community, I'm trying to forward all ESXi-logs to my central log-server, but it does not work. I opened esxi-firewall, set-up Syslog.global.LogHost as tcp://10.0.0.171:514 but all I see on ... See more...
Hi community, I'm trying to forward all ESXi-logs to my central log-server, but it does not work. I opened esxi-firewall, set-up Syslog.global.LogHost as tcp://10.0.0.171:514 but all I see on log-server (10.0.0.171) is: 2013-04-23T19:26:26+00:00 vs1-sys syslog-ng[1552]: Syslog connection accepted; fd='10', client='AF_INET(10.0.0.170:50036)', local='AF_INET(0.0.0.0:514)' 2013-04-23T19:26:26+00:00 vs1-sys syslog-ng[1552]: Invalid frame header; header='' 2013-04-23T19:26:26+00:00 vs1-sys syslog-ng[1552]: Syslog connection closed; fd='10', client='AF_INET(10.0.0.170:50036)', local='AF_INET(0.0.0.0:514)' And all I found on ESXi-server (10.0.0.170) in vobd.log is: 2013-04-23T19:26:48.126Z: [GenericCorrelator] 7587675368us: [vob.user.vmsyslogd.remote.failure] The host "10.0.0.171:514" has become unreachable.  Remote logging to this host has stopped. 2013-04-23T19:26:48.126Z: [UserLevelCorrelator] 7587675368us: [vob.user.vmsyslogd.remote.failure] The host "10.0.0.171:514" has become unreachable.  Remote logging to this host has stopped. 2013-04-23T19:26:48.127Z: [UserLevelCorrelator] 7587675741us: [esx.problem.vmsyslogd.remote.failure] The host "10.0.0.171:514" has become unreachable.  Remote logging to this host has stopped. So where is the problem? I know log-server works, it collects messages from a few servers without any problem. So the problem is somewhere in ESXi-server. BTW I noticed nearly all messages in ESXi-logs (locally saved to files) look strange. The have neither BSD (priority, header, message), nor IETF (header, data, message) format. This might be the problem. Log-server probably does not know how to process such messages...
I see on that screen "Can not load module megaraid_sas: module is already loaded". This imho is not a good idea (to force reloading of driver which is already loaded). Instead of that customise y... See more...
I see on that screen "Can not load module megaraid_sas: module is already loaded". This imho is not a good idea (to force reloading of driver which is already loaded). Instead of that customise your esxi-image (http://www.v-front.de/p/esxi-customizer.html). It has probably older version of megaraid_sas driver, so throw it away from esxi installation image and substitute it with the latest one you download from LSI-webpage.
"...does the NTP client service have to be started or not?..." Yes, it has to be started. "...do we just put in NTP servers..." This is necessary too, ntp-client need to know the source ... See more...
"...does the NTP client service have to be started or not?..." Yes, it has to be started. "...do we just put in NTP servers..." This is necessary too, ntp-client need to know the source of accurate time. "...and set time?..." Not needed. NTP-client will set the correct time itself, after contacting ntp-servers a few times. If you need more info about how ntp is working, please read documentation section on ntp.org.
I think it is possible, with a little tweaking. Do not forget your "raid-drive" will not have the same datastore-name (and path) as the single drive you are using now...
It does not matter which thread of the same core (or cpu) places request on data from memory, because both threads of the same core have the same affinity to particular memory-bank, share the sam... See more...
It does not matter which thread of the same core (or cpu) places request on data from memory, because both threads of the same core have the same affinity to particular memory-bank, share the same L1/2/3 cache, etc. From NUMA point of view they are equal. That's the reason why NUMA does not need to have list of "cpu-threads", it is enough to keep list of cores. You might ask "why is it then not enough to keep just list of cpus"? Because cores inside  of the same cpu might be organised in complex way, clustered, or multi-layered, sharing or not sharing common cache, etc. (i.e. "bulldozer" microarchitecture, 8 cores in 4 clustered modules). Then 2 cores of the same cpu-cluster are for NUMA not the same as 2 cores from different cpu-clusters (but from the same physical cpu)...
Just to make things clear, I suppose with NUMA you mean "non uniform memory access" and with HT "hyperthreading". If it is so, then what has NUMA to do with HT? From NUMA point of view, it does n... See more...
Just to make things clear, I suppose with NUMA you mean "non uniform memory access" and with HT "hyperthreading". If it is so, then what has NUMA to do with HT? From NUMA point of view, it does not matter if real core (cpu) or "hyperthreaded-core" needs data from memory. The only thing which *does* matter: is memory-page requested in "local" (directly accessible) or "non-local" (indirectly accessible) memory bank? HT can have (and probably has) some impact on frequency of "cache misses" (2 threads still share the same amount of cpu-cache) and this might increase number of memory-pages requested, but this effect is very small and outweighed by benefits of running 2 threads in parallel. Or I put it other way: if you do not have problems with HT on UMA (unified memory architecture), you will very probably do not see problems even on NUMA (and vice-versa)...
Well, if you are copying small files, this might be normal. I have seen some test comparing hdd vs ssd where transfer speed while copying small files dropped on hdd even under 1MB/s (depends on f... See more...
Well, if you are copying small files, this might be normal. I have seen some test comparing hdd vs ssd where transfer speed while copying small files dropped on hdd even under 1MB/s (depends on fragmentation, file-size, etc). Tests were done in real environment (no VM) and iirc the hard-drive was "raptor" (one of the fastests hdd). Just for comparison, try to copy single big file. But if you are dealing frequently with small files, maybe it is time to put vm-datastore on ssd...
Please try to be as specific in problem description as you can: What do you mean by "copying 1GB file"? Inside VM? If so, then which OS is VM running (probably some Windows if you use NTFS partit... See more...
Please try to be as specific in problem description as you can: What do you mean by "copying 1GB file"? Inside VM? If so, then which OS is VM running (probably some Windows if you use NTFS partition)? Or in ESXi-console/ssh/ftp/datastore-browser? Copying within one physical drive, or from drive1->drive2? How much memory your VM has (is used for cache & i/o buffers during copying)? Could you check cpu and i/o-load during copying (in VM and hypervisor)? Anyway this "speed" reminds me of old days when some drives per default used only PIO and had to be explicitely instructed to use DMA. This should not be your case, but is maybe worth to check. Also try to remove that cpu-share limit just to see if it has some effect...
I have seen expanders of various form and it is up to you to pick the suitable one. I have been using one in the form of "pcie-card" because I could mount it in an empty slot (apart from that, it... See more...
I have seen expanders of various form and it is up to you to pick the suitable one. I have been using one in the form of "pcie-card" because I could mount it in an empty slot (apart from that, it does not use any pcie-pins). Something like this: http://www.ebay.at/itm/TOP-Chenbro-CK12804-24x-SAS-SATA-Expander-Low-Profile-TOP-/190820907490?pt=Controller&hash=item2c6dcff9e2 There are other types, i.e. one you can mount in empty 2.5" or 3.5"-drive position. And there are also special expanders for which mounting position is already prepared in server-case. Just pick the one with SFF-8087/SFF-8087 connectors, because that's what you have on your LSI raid controller.
That's beause ESXi does not support on-board "fake-raid". AFAIK, only true hardware raid-controllers are supported. They might be sometimes part of motherboard, but this is not the case. "raid"-f... See more...
That's beause ESXi does not support on-board "fake-raid". AFAIK, only true hardware raid-controllers are supported. They might be sometimes part of motherboard, but this is not the case. "raid"-function of your motherboard is done by C602 chipset... My tip: Attach even those two drives you want to use for system to LSI controller and create raid1 there (use expander if 8 ports of your LSI raid controller are not enough). But I think it is maybe "overkill". The drive where you install ESXi is not used normally for anything else, except reading boot-image...
"...Indeed, the VM ESXi 5.1 is the FREE one..." Do not forget free ESXi (hypervisor) is limited to 32GB RAM. No need to buy 64GB. And IIRC, for xeon-cpu you must have ecc-dimms. If 32GB is ... See more...
"...Indeed, the VM ESXi 5.1 is the FREE one..." Do not forget free ESXi (hypervisor) is limited to 32GB RAM. No need to buy 64GB. And IIRC, for xeon-cpu you must have ecc-dimms. If 32GB is not enough for all VMs and you can not (will not) buy i.e. Essential license, you can buy fast SSD and use it for VM-swap.
Probably ESXi does not support your pci-e sata controller. First check Hardware Compatibility List. If you do not find your controller there it is not supported directly by VMware. It could be st... See more...
Probably ESXi does not support your pci-e sata controller. First check Hardware Compatibility List. If you do not find your controller there it is not supported directly by VMware. It could be still supported by hw-manufacturer who produces this controller. Some of them publish vib-drivers for esxi and make them available for download. Finally you can search if someone from user-community created driver. If you do not find anything, chances are low...
Hi community, I decided to move my own internet-server from OS-level virtualization to hypervisor and of course I need remote access for management of my server. So I started googling for some... See more...
Hi community, I decided to move my own internet-server from OS-level virtualization to hypervisor and of course I need remote access for management of my server. So I started googling for some info but what I found was quite frequently "direct exposing of esxi-server to internet is not recommended" opinion. But from security point of view I personally do not see a big difference between OS-level virtualization (i.e. VServer or OpenVZ which I have been running on my internet server for years) and bare-metal hypervisor (correct me if I am wrong), so I would like to try it anyway. My question is: are there any general recommendations how to tighten esxi-configuration so that it could be deployed to the wild and dangerous world full of nasty beasts going by name "internet"? Is here anyone who has esxi-server deployed on internet, administering it over internet? Here is what I have planned to do up to now: I thought about closing all unnecessary ports/services so the only incomming ports opened should be those for ssh and for vsphere-client. Concerning outgoing connection, only dns (udp/tcp 53) and ntp (udp/123 I think) opened. Further I want to move ssh-server from 22 to higher non-standard port, limit incomming ssh-connections only to those comming from selected IP-addresses/segments, set-up new login-name, forbid direct root login and use only keyfiles for authentication (if this all is possinble). I plan to use at minimum 3 interfaces (one dedicated for esxi-management, next one for remote kvm-console, rest for virtual servers). I already talked with admin of server-housing center, he can configure hw-firewall for my server (at least for management- and kvm-interfaces), I just have to let him know what I want. This will act as "1st line of defense", in addition to sw-firewall running on esxi-server. Is there anything else I could? Jarry
You can simply use datastore-browser, but it is quite slow. I have both server and my client-workstation on the same gigabit-lan segment, and I get hardly more than a few MB/s. Unsuitable for cop... See more...
You can simply use datastore-browser, but it is quite slow. I have both server and my client-workstation on the same gigabit-lan segment, and I get hardly more than a few MB/s. Unsuitable for copying big files i.e. cd/dvd iso-mages.You can get a little higher speed with nfs. Set-up nfs server on your client, and attach it to esxi-server as network storage...
I would not over-estimate usefullness of HCL, because it is not reliable at all. In my case I picked motherboard listed as "supported" (without any further notice), but to my dissapointment only ... See more...
I would not over-estimate usefullness of HCL, because it is not reliable at all. In my case I picked motherboard listed as "supported" (without any further notice), but to my dissapointment only one of two on-board network adapters is recognised. I had similar badluck with hw raid-controller (listed as "supported", but no driver for management/health-status). What the heck is such a hardware "compatibility" list good for?
"...I have a LSI MegaRAID SAS 9260-4i without BBU..." I think missing BBU might be cause of your problem. I have much older 8888ELP, I have been using it without BBU and suffered from performa... See more...
"...I have a LSI MegaRAID SAS 9260-4i without BBU..." I think missing BBU might be cause of your problem. I have much older 8888ELP, I have been using it without BBU and suffered from performance too. When I added BBU and enabled write-cache performace skyrocketed. I'd recommend to invest in decent BBU. You can find some cheap 2nd-hand one on ebay... BTW, what disks are you using? I'd recommend to use enterprise-class (or "raid-edition") drives, i.e. drives which are especially designed to run in raid. They have time-limited error recovery, or can even postpone error recovery and sector-relocation for later time, when drive is not being used...