[00:00:47] *** bbrooks_ has joined #Citrix
[00:03:23] *** Splatone_ has joined #Citrix
[00:05:26] *** ScottCochran has joined #Citrix
[00:06:00] <ScottCochran> hey all, how is everyone?  Anything I can help with this evening?
[00:06:19] <Rienzilla> Sam adams, please
[00:06:26] <Rienzilla> :)
[00:08:31] *** bbrooks has quit IRC
[00:08:34] *** deshantm has quit IRC
[00:08:38] *** Appiah has quit IRC
[00:08:38] *** Splatone has quit IRC
[00:09:43] *** Appiah has joined #Citrix
[00:15:31] *** deshantm has joined #Citrix
[00:21:46] *** bbrooks_ has quit IRC
[00:23:45] *** Teck has joined #Citrix
[00:24:37] <ScottCochran> Rienzilla haha
[00:37:27] *** Alasdairrr is now known as AlasAway
[00:37:56] *** Teck has left #Citrix
[00:46:18] <JarianGibson> hey everybody
[00:46:36] *** The_Machine has joined #Citrix
[00:47:12] *** The_Machine70x7 has joined #Citrix
[00:51:28] *** The_Machine has quit IRC
[00:51:33] *** The_Machine70x7 has quit IRC
[01:13:46] *** gblfxt has quit IRC
[01:14:31] *** gblfxt has joined #Citrix
[01:27:03] *** MSilva01 has quit IRC
[01:34:11] *** MSilva01 has joined #Citrix
[02:20:36] *** MSilva01 has quit IRC
[02:26:36] *** MSilva01 has joined #Citrix
[03:45:23] *** waynerr__ has joined #Citrix
[03:49:23] *** lesrar has quit IRC
[03:55:17] *** MSilva01 has quit IRC
[03:58:27] *** kdavy has joined #Citrix
[06:44:30] *** Gio^ has quit IRC
[07:19:12] *** Faithful has quit IRC
[07:57:51] *** ScottCochran has quit IRC
[08:58:19] *** jamesd2 has joined #Citrix
[09:00:00] *** jamesd_laptop has quit IRC
[09:15:10] *** Arrfab has left #Citrix
[10:00:36] *** pesadilla has joined #Citrix
[10:01:17] *** gladier has quit IRC
[10:01:25] *** gladier has joined #Citrix
[10:05:48] *** finnzi has joined #Citrix
[12:42:02] *** ThinkTank has joined #Citrix
[12:43:43] <ThinkTank> Hi, boys and girls. Quick question.. Just upgraded our xenserver lab to 5.6 FP1. Was wondering what iso image to use for Debian Squeeze template or just use "Other Os"
[13:13:18] *** censor has joined #Citrix
[13:13:27] <censor> hi all
[13:13:33] <ThinkTank> hey hey
[13:13:48] <censor> i'm try to work with the XenServer API, but i'm hitting a speed bump...
[13:14:08] <censor> i can clone a template just fine, but provisioning it give me the erro "INVALID_UUID sr"
[13:14:34] <censor> XenServer 5.6.0 FP1 and a PHP XenApi class, pretty simple
[13:14:48] <censor> i can create new VMs from XenCenter and the CLI just fine
[13:14:58] <ThinkTank> what is your parameterlist for the actual command ?
[13:15:21] <censor> i initially wrote a bash script to create the VMs which got send via SSH, should have taken a look at the API first ;-)
[13:15:32] <censor> $xs->VM_provision($vmRef);
[13:15:36] <ThinkTank> hehe
[13:15:46] <censor> where $xs is the object for the XenServer connection
[13:15:58] <censor> let me pastebin my code, sec...
[13:16:01] <ThinkTank> sure
[13:16:47] <censor> http://pastebin.com/SNV4Wiy3
[13:17:33] <ThinkTank> and its failing on $xs->VM_provision($vmRef); ?
[13:19:01] <censor> yup
[13:19:07] <censor> let me get you the log
[13:19:43] <censor> i cut off a couple of fields, should be pretty complete though: http://pastebin.com/gbD991vw
[13:20:21] <censor> i can see that the pool does have a default_SR set...
[13:21:06] <censor> i tried with "VM.copy(vmRef, srRef)", but that results in the same error
[13:22:23] <ThinkTank> what kind of storage is et as your default storage ?
[13:22:35] <ThinkTank> is set*
[13:22:38] <censor> an EqualLogic iSCSI storage
[13:22:46] <censor> two hosts in the pool
[13:22:52] <ThinkTank> and that is the default storage is xenserver?
[13:22:55] <censor> yup
[13:22:57] <ThinkTank> k
[13:23:27] <ThinkTank> you already answered that i see ;O'
[13:23:34] <censor> no prob =)
[13:24:35] <ThinkTank> do you have a session object ?
[13:24:53] <censor> yes, that's being taken care of by the PHP class
[13:25:16] <censor> and i works, as i can clone the template, i see it in XenCenter
[13:25:18] <ThinkTank> i mean a session object for the xenserver api connection
[13:25:24] <censor> yup
[13:25:52] <ThinkTank> in the xenserver api guide it mentions: provision(session, new_vm_ref)
[13:25:53] <censor> the PHP class connects and provides the session on each call
[13:26:40] <censor> here's the class: https://github.com/andygoodwin/PHP-xenapi
[13:27:00] <censor> like i said, the class handles the authentication and session handling
[13:29:37] <ThinkTank> processing the information. Please standby =)
[13:32:03] <censor> sure ;)
[13:37:12] <jduggan_> are you andy goodwin?
[13:37:35] <censor> nope, hehe
[13:37:57] <censor> i'm just using his class, was the first i found
[13:40:06] <censor> or rather, the only...
[13:40:10] <ThinkTank> could you try and invoke SR_get_all_records ?
[13:40:18] <ThinkTank> from your code and print the output
[13:41:13] <censor> sure, sec
[13:42:57] <censor> here you go: http://pastebin.com/EyVAweVk
[13:49:08] <AlasAway> Does anyone have any suggestions regarding: http://pastie.org/private/wvelo1jtd8jnnqi7bfla
[13:50:58] <ThinkTank> censor: is it a 1 x xenserve setup or multiple servers ?
[13:50:59] <AlasAway> That's with XenConvert attempting to import an OVF into XenServer
[13:51:14] <censor> this pool has 2 members
[13:51:21] <AlasAway> I also decided to give KenshoOVF a go, but it complains "WinRM is not installed", despite the fact I just installed it
[13:51:28] <ThinkTank> and both has access to the storage backend ?
[13:51:33] <censor> yes
[13:51:48] <ThinkTank> is it a production environment ?
[13:51:56] <censor> unfortunately, yes
[13:52:11] <ThinkTank> ah
[13:52:29] <ThinkTank> well.. i would like to try setting local disks as default storage backend and run your script
[13:52:38] <censor> i do have another low-priority standalone box where i can try that
[13:52:46] <censor> hang on
[13:52:51] <ThinkTank> to check if the problem is to provision to your san box
[13:53:29] <censor> nope, same error =(
[13:53:50] <ThinkTank> k
[13:53:50] *** kprojects has joined #Citrix
[13:57:06] <ThinkTank> censor: could to also print out the output of $vmRef ?
[13:57:13] <censor> sure
[13:58:47] <censor> http://pastebin.com/Lk5xTEqg
[13:59:34] <censor> this i from the cloned VM, then VM.get_record(vmRef)
[13:59:38] <censor> *is
[13:59:57] <censor> at the bottom you can see the error from VM_provision($vmRef)
[14:01:02] <censor> brb, meatspace meeting
[14:01:15] <censor> but thanks for your support so far, i'm really at loss here =)
[14:01:22] <ThinkTank> did you make the template yourself ?
[14:01:35] <censor> nope, it's the stock one from 5.6.0 FP1
[14:02:01] <ThinkTank> ah
[14:02:11] <ThinkTank> i don't believe you can use the clone/provision that way
[14:02:28] <ThinkTank> the stock template isent "really a template" :P
[14:02:33] <ThinkTank> its a installer template
[14:02:49] <censor> yeah well wo do i setup a new VM then?
[14:03:01] <censor> according to the docs, that's the way to go... i think i read
[14:03:19] <censor> anyways, really gotta hurry, i'll be back in a few - thanks! =)
[14:07:36] <AlasAway> Has anyone here done a P2V of a Windows Server 2008 box?
[14:08:11] <ThinkTank> tried on win2003 server.. crashed and burned :P but it said to work well in 2008
[14:10:07] <AlasAway> Using XenConvert?
[14:10:47] <ThinkTank> aye
[14:10:55] <ThinkTank> the xenconvert client locked up
[14:10:58] <ThinkTank> on 2003 server
[14:11:09] <ThinkTank> haven't tried on 2008 thought
[14:11:18] <ThinkTank> but i was reading on forums that 2008 was alot easier
[14:15:09] <AlasAway> So far I can't get it to actually spit the image into XenServer itself
[14:15:36] <AlasAway> i'm just trying as an OVA so XenCenter can do the import
[14:15:55] <ThinkTank> so you get the OVA image ?
[14:16:22] <AlasAway> yeah
[14:16:26] <ThinkTank> If you have the OVA image there is actually to ways to go around
[14:16:47] <ThinkTank> either put the OVA to a shared storage (nfs)
[14:17:11] <ThinkTank> this way you can mount up the storage from the xenserver console and use xe vm-import
[14:17:29] <AlasAway> that sounds better
[14:18:08] <ThinkTank> the alternative is to use xenCenter.. but its really slow since it uses SSH to send the data to xenserver
[14:18:26] <AlasAway> *nods* - I can do NFS quite easily so I'll go that route
[14:18:37] <ThinkTank> thats best
[14:18:54] <AlasAway> Thanks for the assistance/suggestions :)
[14:18:58] <ThinkTank> no problem
[14:19:07] <ThinkTank> its nice to have a nfs server for cases like this
[14:19:15] <ThinkTank> you also have the xe vm-export command
[14:19:31] <ThinkTank> it takes a vm and spits it out to a location as xva
[14:19:47] <AlasAway> We should probably use these tools far more often
[14:19:50] <ThinkTank> really good for moving vm's between clusters
[14:19:54] <ThinkTank> indeed
[14:20:07] <ThinkTank> XenCenter is a good tool no doubt about that
[14:20:16] <ThinkTank> but i just dont use it for operations like this
[14:20:48] <AlasAway> Makes perfect sense
[14:20:52] <AlasAway> :)
[14:20:55] <ThinkTank> yep =)
[14:21:30] <ThinkTank> Give a shoutout if you arrive at any pitfalls on the way. Shouldnt be any thought
[14:22:22] <ThinkTank> When doing p2v and migration its always nice to have a dartboard with murphys law on it
[14:22:23] <ThinkTank> :P
[14:23:59] <Rienzilla> haha
[14:24:09] <ThinkTank> =)
[14:24:30] <Rienzilla> true for most migrations of any type really :)
[14:24:34] <ThinkTank> indeed
[14:24:53] <Rienzilla> no matter how well you plan, something will go wrong :)
[14:24:59] <ThinkTank> i prefer http://www.thinkgeek.com/geektoys/warfare/8bc4/ thought
[14:25:41] <Rienzilla> lol i actually have that
[14:25:47] <Rienzilla> without the laser
[14:26:34] <ThinkTank> i have considered making a arduino and writing a own targeting system for it.
[14:27:23] <ThinkTank> nerd rage
[14:27:51] <Rienzilla> hehe
[14:27:53] <Rienzilla> awesome
[14:29:20] <ThinkTank> i wish the they could cure sleep
[14:29:29] <ThinkTank> then i would have time for stuff like that
[14:33:04] * ThinkTank afk grabs a cup of coffee
[14:36:41] <AlasAway> Hmm, my "Appliance Import" optino is missing from XenServer
[14:36:46] <AlasAway> XenSCenter
[14:36:48] <AlasAway> er
[14:36:53] <AlasAway> XenCenter
[14:39:12] <ThinkTank> did you want to imort in xencenter anyways ?
[14:40:47] <ThinkTank> lets say you have a cluster of 2 xenservers. You right click the server you wish to import it to. IT should be a option there called import
[14:44:38] <AlasAway> http://community.citrix.com/display/ocb/2010/06/09/XenServer+5.6+OVF+Appliance+Wizards
[14:44:46] <AlasAway> That seems to contradict the documentation
[14:44:53] <AlasAway> I guess its just "import" now
[14:45:41] <ThinkTank> propably
[14:45:52] <ThinkTank> so you exported your live machine to ovf ?
[15:21:37] *** frogstarr78 has quit IRC
[15:22:01] *** frogstarr78 has joined #Citrix
[15:28:38] *** arneh has joined #Citrix
[15:43:23] *** arneh has quit IRC
[16:01:14] <censor> ThinkTank: i'm back! so regarding the cloning of an installer template, i think that's how you're supposed to do it - i don't want to literally clone an existing VM, but rather clone the "empty" template, provision it with VDIs, and then install from scratch
[16:01:31] <ThinkTank> ah ok
[16:01:53] <ThinkTank> it seems that you need to do some more "manual" work if you wanna do the plain template
[16:01:53] <censor> just saw the privmsg, what did you find?
[16:02:00] <ThinkTank> hold on
[16:02:04] <ThinkTank> gonna find the tab
[16:03:45] <ThinkTank> here it is
[16:04:36] <censor> btw, here's where i read how to install a new VM via the API (although it's from 4.0.1, i think the method hasn't changed that much): http://docs.vmd.citrix.com/XenServer/4.0.1/sdk/ch03.html#id2483076
[16:04:55] <ThinkTank> well what i did is checked the api docs
[16:04:58] <ThinkTank> for examples
[16:05:05] <ThinkTank> i found a java example
[16:05:13] <ThinkTank> http://community.citrix.com/display/xs/Download+SDKs
[16:05:27] <ThinkTank> xenserverjava samples
[16:05:40] <ThinkTank> unzip and check out CreateVM.java
[16:05:42] *** The_Machine has joined #Citrix
[16:06:08] <ThinkTank> basically you need to "find" the defaultSR
[16:06:19] <ThinkTank> and setup all the network stuff
[16:06:26] <ThinkTank> its well documented
[16:07:28] <censor> yea, i was going to take care of the network, but i needed the VDI first - and the template says it already provides an VDI
[16:07:44] <censor> like if i install it from the CLI, i don't have to do that either
[16:10:14] <ThinkTank> it just seems that the xapi dosen't know which SR to use. Normally when you do clone/provision through xapi it chooses the same SR as the template is on. Since the "standard templates" isent really on any SR (Its bundled somewhere) it wont be able to find the SR
[16:10:28] <ThinkTank> this is kinda speculation thought
[16:11:11] <censor> hmm, kinda makes sense, but i'm wondering how the CLI does it then
[16:12:55] <ThinkTank> i think its a question if xapi wraps around cli or cli wraps around xapi
[16:13:25] <censor> i'd say the cli is a xapi wrapper
[16:13:35] <censor> as granular as it is...
[16:13:52] <censor> i'll give the manual way a shot
[16:31:57] <jduggan_> kdavy ?
[17:00:18] *** rev78 has joined #Citrix
[17:11:25] <ThinkTank> censor: How is it going ?
[17:12:55] *** Gio^ has joined #Citrix
[17:13:05] <censor> struggling with the VIFs =)
[17:14:09] <censor> the java classes are happy with supplying 4 parameters to the xapi, but i get complains about missing fields
[17:14:54] <deshantm> censor: just so you know there is a #xen-api channel too
[17:15:13] <censor> oh, good to know, thanks!
[17:33:24] *** MSilva01 has joined #Citrix
[17:33:39] <MSilva01> morning ...
[17:51:21] *** kdavy_ has joined #Citrix
[17:51:30] <kdavy_> morning
[17:56:08] <jduggan_> kdavy: what was the brother model>
[17:56:09] <jduggan_> ?
[17:56:54] <kdavy_> jduggan_: MFC-8860DN
[18:01:29] <jduggan_> kdavy_: not that i dont trust you - but looking at the manual i don't see the FTP feature?
[18:04:20] <kdavy_> jduggan_: http://welcome.solutions.brother.com/bsc/public_s/id/us/us/en/monolasermfc/mfc8860dn_all/spec/index.html
[18:04:29] <kdavy_> Scan to FTP 	Yes
[18:04:57] <kdavy_> Scan to Email Server 	Yes
[18:33:46] *** tom_wurm has quit IRC
[18:38:21] <kdavy_> ok wtf. i have a broken SR in xenserver
[18:38:25] <kdavy_> no idea how to fix it
[18:38:43] <kdavy_> got one VM stuck on it that i need to move
[18:39:10] <kdavy_> when i try to do anything to the VM - start, move, etc - i get a "VDI is not available" error
[18:40:43] <kdavy_> Error code: SR_BACKEND_FAILURE_46 Error parameters: , The VDI is not available [opterr=Interrupted operation detected on this VDI, scan SR first to trigger auto-repair]
[18:42:03] <kdavy_> then when i do sr-scan, i get "There was an SR backend failure." with a bunch of tracebacks
[18:43:18] <kdavy_>  /var/log/SMlog shows the following:
[18:43:38] <kdavy_> ['/usr/sbin/vhd-util', 'modify', '--debug', '-s', '8388608', '-n', '/dev/VG_XenStorage-582ff1c9-8ee7-6de4-9a13-cd8595d9671c/VHD-3f7fc36e-7ffb-4c0c-98ee-bab34e8c8285'] - FAILED: (rc 22) stdout: 'failed to set physical size to 8388608: -22
[18:54:26] *** OmNomDePlume has joined #Citrix
[19:11:02] *** OmNomDePlume has quit IRC
[19:11:26] *** OmNomDePlume has joined #Citrix
[19:17:04] *** tom_wurm has joined #Citrix
[19:24:24] *** SpyderZ has joined #Citrix
[19:35:34] *** SpyderZ has quit IRC
[19:38:54] *** scsinutz has joined #Citrix
[19:39:15] *** scsinutz has left #Citrix
[20:09:21] <Splatone_> kdavy_: any luck?
[20:09:47] <kdavy_> Splatone_: nope
[20:09:49] <kdavy_> :(
[20:10:16] <Splatone_> kdavy_: Is everything hung up on the hosts?  Ie the task queue?
[20:10:37] <kdavy_> what sucks most is, the VM itself is holding no data that can't be recreated automatically, but recreating the data will take 2 weeks - it is a full text index of 500Gb worth of email
[20:11:16] <kdavy_> Splatone_: nothing is hung up. only this SR is giving issues, and this is the only VM on the SR
[20:11:45] <Splatone_> kdavy_: anything different with the SR?
[20:12:46] <kdavy_> nope, aside from the fact that it is the oldest SR on the system - was created on XS 5.5
[20:13:52] *** The_Machine has quit IRC
[20:15:56] <kdavy_> actually i take that back. there are a few other VMs on it, but they run just fine. test VMs
[20:16:37] <Splatone_> kdavy_: my guess is that the host thinks the vdi is locked.
[20:17:49] <kdavy_> nope, it isnt locked
[20:19:29] <kreign> kdavy, which vendor is it that you get all your supermicro kit from?
[20:20:50] <kdavy_> kreign, wiredzone.com
[20:21:53] <kreign> kdavy, they typically just have kit or do they have all the misc. adapters like the supermicro backwards riser cards?
[20:22:15] <kdavy_> kreign: they have pretty much everything
[20:25:36] *** The_Machine has joined #Citrix
[20:27:11] <kdavy_> Splatone_: any other idea what it could be?
[20:27:37] <kdavy_> i've googled around and found nothing useful
[20:29:45] <kdavy_> oh, btw. this was the only VM where i had daily scheduled snapshots enabled via the Protection and Recovery feature. I figure the looooong chain of snapshots might have something to do with this
[20:36:24] <kdavy_> good thing i have an extra xenserver and my SAN allows snapshots. i'll play with a copy of the SR
[20:40:53] <kdavy_> but it's shit like this, Citrix... why does this never happen on VMware?
[20:41:25] <kdavy_> oh, wait, i know why. VMware doesn't use LVM
[20:41:30] *** The_Machine has quit IRC
[20:41:47] *** kprojects has quit IRC
[20:42:13] <kreign> kdavy, ...
[20:42:16] <kreign> oh
[20:42:24] <kreign> confusion results when you read the fuckiing buffer backwards
[20:42:36] <kreign> kdavy, LVM sucks nutsack by choice.
[20:45:07] <kdavy_> indeed
[20:47:13] <kdavy_> sweeeet! found a manual SAN snapshot of the volume from last weekend - completely forgot i made it. 2 hours and it would've been deleted. that should have the SR in a non-corrupt state
[20:48:09] <kdavy_> which says something about my trust in XenServer - any time i do a medium to major change, i snapshot all SRs on the SAN
[20:49:48] *** blood has joined #Citrix
[20:50:25] <blood> If I extend one of my centos virtual disks, is there a way to extend the logical volume with LVM?
[20:55:38] *** kprojects has joined #Citrix
[20:57:05] <blood> basically i'm trying to resize an LVM partition
[21:02:20] *** kprojects has quit IRC
[21:04:29] *** finnzi has quit IRC
[21:06:39] *** AlasAway is now known as Alasdairrr
[21:13:37] *** deshantm_xenorg has joined #Citrix
[21:14:03] <kdavy_> blood: theoretically yes
[21:14:31] <kdavy_> though i don't know how and i would advise against trying
[21:18:12] *** Alasdairrr has quit IRC
[21:18:18] *** Alasdairrr has joined #Citrix
[21:18:42] <blood> Well, i'm thinking I just create an extended partition then each time I want to add more space I can just add a logical partition inside my extended. Then I would pvcreate the newly created partition and extend my logical group. make sense?
[21:20:51] <kdavy_> blood, makes sense theoretically
[21:21:01] <kdavy_> i try to deal with LVM as little as i can
[21:31:44] *** deshantm_xenorg has quit IRC
[21:33:03] *** slim_ has joined #Citrix
[21:33:55] *** pesadilla has quit IRC
[21:43:34] *** deshantm_xenorg has joined #Citrix
[21:51:14] <slim_> hi
[21:52:07] <kreign> blood, the LVM resize isn't theoretical - making it bigger isn't such an issue. i just don't know how XS will like it.
[21:52:22] <kreign> it's a logical volume manager
[21:52:27] <kreign> you can have it span a dozen arrays if you'd like
[21:52:32] <kreign> i just wouldn't.
[21:52:41] <blood> yea we just want to use one virtual disk and increase it as needed
[21:52:42] <kreign> (use it at all, if I could avoid it)
[21:53:01] <kreign> blood, if that's what's on your plate i suggest you look at more scalable options.
[21:53:07] <kreign> and uniform
[21:53:18] <blood> so basically don't use LVM
[21:53:40] <blood> I mean we could simply create a new virtual disk and create a new physical volume to toss into our LVM every time
[21:54:19] <blood> goal is to make future resizes as easy as possible
[21:55:46] <blood> i'll dive into this more tomorrow
[21:55:47] <blood> take care
[21:55:51] *** blood has quit IRC
[21:57:41] *** slim_ has quit IRC
[22:00:05] *** archstanton43 has joined #Citrix
[22:10:32] *** pesadilla has joined #Citrix
[22:18:09] <kdavy_> hmm so i get an SR Backend Failure even when i try to attach a storage-level clone of that SR to another host. something is seriously fucked there
[22:19:20] <kreign> kdavy, oh that's wonderful.
[22:19:41] <kreign> kdavy, every time I see an LVM issue I don't feel so bad about using a beta filesystem  implementation in production.
[22:20:12] <kreign> I expect a -reasonable- amount of Magic in my IT tools, god damn it.
[22:21:03] <kdavy_> kreign, maybe you can tell what's wrong by this log excerpt? http://pastebin.com/p6ULCWpm
[22:21:42] <kdavy_> that's everything from the SR attach procedure. the LVM volume of my stuck virtual machine is the one starting with 3f7f
[22:22:06] <kdavy_> er, well, the VHD-3f7fc36e-7ffb-4c0c-98ee-bab34e8c8285
[22:23:29] <kreign> hold up
[22:23:33] <kreign> the bitches be at the door
[22:23:39] <kreign> argg
[22:23:44] <kdavy_> k
[22:24:01] <kreign> I really hate "everything is hidden by UUIDs"
[22:24:12] <kreign> because labeling things like a human is Hard
[22:24:29] <kreign> I can't wait for VIFs with IPv6 addresses.
[22:25:05] <kdavy_> kreign: humans and uniqueness don't work well together. just try asking someone to say a random word - it'll bring them into a stupor for a couple seconds
[22:25:48] <kreign> kdavy, any idea what the deal is with that dd if=/dev/zero instantiation  w/ the large seek?
[22:25:52] <kreign> (what's it doing)?
[22:26:06] <kreign> kdavy, my random word is 'blueberry'
[22:26:11] <kreign> but then I'm quick on my feet
[22:26:15] <kreign> (and consistent)
[22:26:16] <kreign> :P
[22:26:42] <kreign> kdavy, I"m not saying use the human readable name instead of the uuid, i'm saying use it in addition.
[22:26:53] <kreign> eg.: xe vm-list should output a human-readable list. :P
[22:27:06] <kreign> (the cli tool syntax is pretty descent unless you're doing something a bit more complex, though)
[22:27:18] <kdavy_> kreign: could be quiescing of snapshots, but i have no clue otherwise
[22:27:52] <kdavy_> the large seek corresponds to the size of the vdisk - 50 gigs
[22:28:11] <kreign> how big is this pv supposed to be?
[22:28:12] <kreign> lvm
[22:28:21] <kreign> oh
[22:28:48] <kdavy_> pv as in the storage repository or the virtual disk?
[22:29:21] <kreign> kdavy, looks like vhd-util is fucking up.
[22:29:29] <kreign> edits with status -22
[22:29:44] <kdavy_> yeah, but why is it fucking up and how do i fix it? :)
[22:30:42] <kreign> tho i think that's not a normal exit status
[22:30:51] <kreign> seems a bit high, seem to recall some sort of artificial limit
[22:31:27] <kreign> kdavy, run pvck on the physical volume
[22:32:11] <kreign> and maybe pvscan?
[22:32:29] <kreign> I dunno, the XS errors mean fuckall to me, tbh
[22:32:55] <kreign> seems like you're probably correct; it's looking for the start of the vhd inside the pv and not finding it, I Think
[22:33:28] <kreign> and not the whole thing
[22:33:48] <kreign> I don't use snapshots.
[22:33:52] <kreign> sorry, can't help ya
[22:33:56] <kreign> (much more than that)
[22:34:25] <kreign> if I could do anything like that with XS, I'd pick for vhds on NFS or raw blocks on iSCSI
[22:34:36] <kdavy_> mm, pvck running now
[22:35:03] <kdavy_> hopefully on the correct volume
[22:35:11] <kreign> lvm is really the wrong way to implement the kind of frequent/large snapshots you'd get out of XS
[22:35:21] <kreign> pvck shouldn't be destructive.
[22:35:48] <kreign> kdavy, lspv and pvs are also your friends
[22:35:51] <kdavy_> kreign: it shouldn't be destructive unless the volume is attached to something else at the same time
[22:36:05] <kreign> right, i realized what you meant afterwards :P
[22:36:05] <kdavy_> anyway the xenserver just kicked me out and now i cant connect to it. lol
[22:36:14] <kreign> so, bad choice.
[22:36:16] <kreign> :|
[22:36:28] <kreign> or rather, 'wrong'
[22:37:31] <kdavy_> i can still ssh though, no pvck process showing in 'top' - does that mean it's done already?
[22:38:24] <kdavy_> aha, xapi crashed
[22:42:15] <kdavy_> pvck thinks for ~30 seconds, then says "Killed" and quits. what does that mean?
[22:42:45] <kdavy_> kreign: ^
[22:43:36] <kreign> kdavy, i'm guessing maybe a contentous kernel lock with xapi or something XS is doing.
[22:43:42] <kreign> kdavy, or possibly it's quite a bit borked.
[22:43:49] <kdavy_> nah, xapi is already dead
[22:43:56] <kreign> kdavy, you sure you're on the write pv?
[22:43:59] <kreign> right
[22:44:22] <kdavy_> yeah, i'm positive
[22:45:22] <kdavy_> unless i'm running pvck wrong. i should just give it the /dev/sd* as argument, right?
[22:45:43] <kreign> one sec
[23:15:27] *** OmNomDePlume has quit IRC
[23:17:20] <kdavy_> kreign: ok i recovered the VM. thank god for that replay from a few days ago
[23:17:51] *** archstanton43 has quit IRC
[23:24:04] <kreign> kdavy, lol what'd you have to do to get it back?
[23:27:43] <kdavy_> kreign: i had a SAN snapshot of the volume from a few days ago
[23:27:58] <kdavy_> from before it got corrupted
[23:30:54] <kreign> cool.
[23:32:47] <kdavy_> so yeah, my Nexenta is in full production now. all 550 xenapp users on it, and it's only 10% loaded from an IOP standpoint
[23:34:11] <kdavy_> and i didn't even attach any solid state disks to the volumes - i want to establish a baseline without solid state and then see how much it changes
[23:34:55] <kreign> kdavy, is that SSD IOPS or or spindle?
[23:35:03] <kreign> oh
[23:35:04] <kreign> nevermind
[23:35:07] <kdavy_> kreign: right now it's all spindle IOPs
[23:35:08] <kreign> *reading comprehensioN*
[23:35:14] <kreign> yeah
[23:35:19] <kreign> you're not goign to max that out
[23:35:37] <kreign> kdavy, was going to say, you actually have your spindles (iirc) a bit under provisioned for 550 users.
[23:35:47] <kreign> we've got 120 and waaay fewer spindles w/o issue
[23:35:48] <kdavy_> no, not anytime soon
[23:36:11] <kdavy_> yeah, way underprovisioned
[23:36:12] <kdavy_> but i
[23:36:23] <kreign> kdavy, IMO that's the only way to do it.
[23:36:28] <kdavy_> i'm going to put Exchange 2010 and a few SQL servers on it as well
[23:36:54] <kreign> IMO even with 5900 RPM disks (eg. 30% utilized) you're going to save a LOT of money
[23:36:58] <kreign> ah
[23:37:01] <kreign> see, that'll be evil
[23:37:05] <kreign> moreso the exchange
[23:37:13] <kdavy_> kreign: how many spindles do you have for 120 users?
[23:37:46] <kreign> kdavy, 12
[23:37:46] <kreign> I think
[23:37:46] <kdavy_> hehe. i have 120
[23:37:50] <kreign> :-o
[23:37:54] <kdavy_> 10k rpm, too
[23:38:03] <kreign> yeah these are 12 10k
[23:38:09] <kreign> actually 16 now that I think a bit better :)
[23:38:13] <kreign> still
[23:38:18] <kdavy_> plus SSDs
[23:38:23] <kreign> no SSDs for me.
[23:38:36] <kreign> my XenServer setup is retarded.
[23:38:39] <kdavy_> SSDs are fun
[23:39:32] <kdavy_> but yeah with the current setup i'll run out of FC link throughput way before i run out of IOPs
[23:41:03] <kreign> kdavy, out of curiosity are you familiar w/ zfs encryption (ie is it possible)?
[23:41:06] <kreign> i've not readily looked into it
[23:41:19] <kreign> (though I'd actually wager that zfs is as close to encryption as you'll get without the pool list)
[23:41:27] <kreign> in terms of filesystems
[23:41:46] <kreign> kdavy, that's almost good IMO.
[23:41:50] <kreign> kdavy, running out of IOPS sucks.
[23:42:36] <kdavy_> 5 XenServers -> 10x2Gb FC -> 2x Sanbox 5200 -> 4x2Gb FC -> 2x Nexenta controllers -> 2x4Gb FC -> 2x Sanbox 5600 -> 12x2Gb FC -> 6x FC JBODs
[23:43:19] <kdavy_> plus Nexenta Controllers -> 16x6Gbps SAS -> 2x SAS JBODs
[23:44:21] <kdavy_> zfs has no encryption yet, it's being worked on
[23:48:22] <kdavy_> kreign: i know, i ran out of IOPS on the Compellent, which is why I built the Nexenta
[23:52:36] <kdavy_> now I'm spamming Nexenta support with bugs in their GUI and plugins :) if i have gold support, might as well get stuff fixed
[23:52:57] <kdavy_> got like 6 incidents open
[23:59:08] <Alasdairrr> Hi all - i'm getting "msg: INTERNAL_ERROR: [ Failure("int_of_string") ]" when trying to import an ovf on xenserver 5.6
[23:59:52] <kreign> kdavy, I'd much rather hit fabric limits than storage limits. fabric is, architectually, easier to replace





top