Switch to DuckDuckGo Search
   October 14, 2013  
< | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | >

bottom
[02:09:06] <fxhp> scheduling snapshots could be a neat feature
[02:09:25] <fxhp> just testing snapshots, super awesome and super fast
[02:09:43] <fxhp> still don't know how to do it on CLI
[02:22:32] <trentster> fxhp: yeah, I agree it would be an awesome feature. I have brought it up in the past.
[02:22:50] <trentster> but you can schedule snapshots cron and fifoadm
[02:23:04] <trentster> but I agree it would be great to be able to do it per vm from within the UI
[02:39:45] <fxhp> I don't know the project-fifo code-base enough so I'm not sure how much work that would be.
[02:40:33] <fxhp> trentster: do you have an example cron + fifoadm snapshot script?
[02:40:48] <fxhp> I'm a complete ZFS newb
[02:45:13] <fxhp> testing cloning now
[02:50:49] <trentster> fxhp: sorry, no I am not currently using fifoadm for snapshots I am using automated zsnapper
[02:51:05] <trentster> I will add an article to my blog soon about using it.
[02:51:42] <trentster> I also plan to look at getting it to work with fifoadm so we can see teh snapshots within jingles ui. but currently my node.js fu is not strong.
[02:56:08] * fxhp nods
[02:56:39] * fxhp does some work with python (Pyramid/flask)
[02:57:15] <fxhp> So I tried to create a new VM from a snapshot, but it just sits at creating
[02:58:10] <fxhp> never shows up in GZ
[02:58:18] <fxhp> using vmadm
[03:00:43] <fxhp> I think I skipped this step "sm-prepare-image"
[03:03:12] <trentster> fxhp: sm-prepare image is vital.
[03:03:16] <trentster> you must do it.
[03:03:17] * fxhp nods
[03:03:26] <fxhp> found an old IRC log
[03:03:27] <trentster> steps are:
[03:03:30] <trentster> create vm.
[03:03:37] <trentster> install software and apps etc and settings
[03:03:50] <trentster> then last step = sm-prepare-image
[03:04:07] <trentster> then from within fifo/jingles
[03:04:41] <trentster> you can click i=on the vm then create a new dataset from it, it will get pulled into the built in dataset server and then you can deploy new vms from it.
[03:05:53] <fxhp> ok, is sm-prepare-image run on the hypervisor ?
[03:06:40] <trentster> fxhp: https://www.monosnap.com/image/OgVctCg1nnkk7coL0uSQ7dvI5
[03:07:06] <trentster> no sm-prepare-image is run from within the machines zone.
[03:07:22] <trentster> This is assuming you creating a zone dataset? and not kvm?
[03:07:30] <fxhp> /zones/cdc5e1eb-db06-44ea-bd53-2c331bff842e/root/usr/img/test/mk-custom-image
[03:07:33] <trentster> for kvm, you dont use sm-prepare-image
[03:08:02] <trentster> fxhp: zlogin cdc5e1eb-db06-44ea-bd53-2c331bff842e
[03:08:05] <trentster> then:
[03:08:09] <trentster> sm-prepare-image
[03:08:31] <fxhp> oh i c
[03:08:42] <fxhp> so I need to login to the zone (that holds the KVM)
[03:09:47] <trentster> no, if its KVM machine, you dont use sm-prepare-image, thats only for zone machines.
[03:10:06] <fxhp> i c
[03:10:19] <trentster> for kvm, just shut down the machine once its ready, and follow the same steps within fifo ui to create a dataset
[03:10:20] <fxhp> yes, right now I'm messing with KVM
[03:10:47] <fxhp> ok that was the issue, I didn't shut it off before creating a cloned dataset
[03:10:54] <fxhp> let me test that out
[03:15:20] <fxhp> still acting the same but if history is any indicator, I
[03:15:32] <fxhp> most certainly did something out of order
[03:15:42] <fxhp> (at least in regards to fifo)
[03:18:56] <trentster> fxhp: you creating a dataset from a snapshot within fifo
[03:19:04] <trentster> you shut down the vm?
[03:19:08] <trentster> create snapshot?
[03:19:16] <trentster> then in fifo create dataset from snapshot?
[03:19:30] <fxhp> yes.
[03:19:43] <fxhp> It seems pretty straight forward
[03:20:12] <fxhp> but when I attempt to use the new dataset for a new KVM guest, it stays in the creating state
[03:20:18] <fxhp> and never finishes
[03:20:32] <trentster> fxhp: aaah how big is your kvm vm?
[03:20:47] <trentster> in the GZ of your node do a:
[03:20:50] <fxhp> I think 16G and 20G
[03:21:22] <fxhp> (I set 20G but the KVM img asks for 16G so two devices show up )
[03:21:36] <trentster> tail -f /var/log/chunter/debug.log
[03:21:41] <fxhp> k
[03:22:20] <trentster> what happens is that the first time you deploy a vm to a node, and the base dataset does not exist on that node, it has to be sent there via fifo first, in chunks
[03:22:54] <trentster> also when you first create a dataset from a snapshot it also has to be pushed via riak into a datastore in redundent chunks. Takes awhile as well.
[03:23:20] <fxhp> I see
[03:23:50] <fxhp> Ok, going to run the tail
[03:23:56] <trentster> and be aware the dataset server is also designed for redundency, so it create 3 chunks of data for the dataset, which if you have multiple fifo services distributed these 3x chunks will be distributed in a redundent fashion.
[03:24:03] <fxhp> should I delete the dataset and create it again?
[03:24:32] <trentster> if you only have a single fifo instance (which you have) all 3 chunks go to same fifo zone, so you neeed 3X the space for teh dataset.
[03:25:28] <fxhp> yeah, woah
[03:25:34] <trentster> make sure that when you first create the dataset within fifo. you go look at the datasets page and wait for a "tick" to show next to teh dataset name in the list. That indicates its finished. You may need to relaod the page
[03:25:59] <fxhp> I'm only playing with one hypervisor
[03:26:03] <fxhp> one fifo
[03:26:20] <fxhp> I did see the tick, but it was really quick, almost seems like its too quick
[03:26:26] <trentster> — in regards to teh 3x space, make sure you have assigned enough "quota" space to your fifo zone to handle all your datasets 3X
[03:26:42] <trentster> fxhp: what version of fifo you running?
[03:28:37] <fxhp> latest I think,
[03:28:43] <fxhp> I documented my environment here http://jira.project-fifo.net/browse/JIN-69
[03:28:45] <fxhp> in that ticket
[03:29:54] <fxhp> so the original KVM dataset is only 500M, are you talking about that size ?
[03:30:10] <fxhp> so the fifo zone would end up storing that as 1.5G ?
[03:30:29] <Licenser> a bit more actually given management overhead
[03:32:23] <fxhp> Ok but when we are talking size, we are talking dataset size, not the size the VM thinks it has
[03:33:26] <trentster> Licenser: do you currently have a "current" master release available somewhere to verify kvm dataset creation is still working?
[03:33:58] <Licenser> trentster nope I'm using dev everywhere ;)
[03:34:27] <trentster> yeah, me too.
[03:34:29] <fxhp> trentster - I could open my fifo up to the internet
[03:34:44] <fxhp> its all just test data right now
[03:35:09] <trentster> fxhp: best to upload fifo logs to a ticket or dropbox if Licenser has time to have a quick look?
[03:35:47] <trentster> fxhp: to be honest I would just run latest dev, its pretty darn stable. Licenser do you concur?
[03:36:05] <Licenser> yup
[03:36:35] <fxhp> I could try that too, fifo / smartos makes it so simple
[03:36:58] <fxhp> I'll test that next
[03:37:08] * fxhp is now QA of fifo-project
[03:37:11] <fxhp> haha
[03:37:21] <Licenser> nah thats trentsters job
[03:37:27] <fxhp> ok good
[03:37:45] <Licenser> he'll cry if you do take his job :P
[03:37:48] <fxhp> lol
[03:37:55] <fxhp> I don't need another job
[03:40:38] <trentster> fxhp: the pays not great :-P
[03:40:57] <fxhp> trentster: I'm going to stop my "master" build of fifo and create a new "dev" build
[03:41:23] <fxhp> trentster: the only part that might collide is chunter
[03:42:05] <fxhp> but it looks like both guides use latest
[03:42:17] <fxhp> so nm, I don't have to mess with chunter
[03:43:34] <trentster> you need chunter dev version as well.
[03:43:52] <trentster> as per the dev install guide on the wiki
[03:43:58] <fxhp> oh you are right
[03:44:14] <fxhp> I read "latest" on both and thought they were the same.
[03:58:21] <fxhp> http://pad.yohdah.com/176/066f295c-ca77-4698-b826-072572a0555a
[04:00:27] <fxhp> so yeah, its crashing for me on master/stable
[04:00:46] <fxhp> I'm going to watch this movie and then install dev
[04:01:14] <trentster> fxhp: cool, pls let us know if it works on dev for you.
[04:28:31] <fxhp> trentster: step 1 of both master and dev use different images
[04:28:48] <trentster> ?
[04:28:51] <fxhp> however the grep string is the same
[04:29:17] <fxhp> grep fdea will not match any chars in the dev image
[04:30:31] <trentster> fxhp: good catch...
[04:30:55] <trentster> fixing now
[04:30:56] <fxhp> should they both be using the same image?
[04:32:09] <trentster> nah, in dev we using a newer smartos base dataset, thats what we test with.
[04:32:29] <trentster> In all theory they should all work, but we prefer to just put steps in we have tested.
[04:32:40] <fxhp> i c
[04:33:56] <trentster> fxhp: you can also use 13.2.1 base 64 if you like
[06:11:36] <trentster> fxhp: Did you emerge victorious?
[06:26:33] <fxhp> trentster: messed up a bit
[06:27:00] <fxhp> purged both my fifo-master and fifo-dev, starting a fresh zone with fifo-dev
[06:30:54] <fxhp> Nope same error exists on dev release
[06:30:57] <fxhp> 2013-10-14 04:29:16.558 [error] <0.5455.0> gen_fsm {vm,<<101,102,100,54,99,51,48,100,45,56,55,48,54,45,52,56,57,99,45,98,98,102,48,45,49,55,56,57,102,102,48,55,55,99,55,48>>} in state creating terminated with reason: no match of right hand value {error,enoent} in chunter_zoneparser:convert/2 line 105
[06:30:58] <fxhp> 2013-10-14 04:29:16.559 [error] <0.5455.0> CRASH REPORT Process <0.5455.0> with 0 neighbours exited with reason: no match of right hand value {error,enoent} in chunter_zoneparser:convert/2 line 105 in gen_fsm:terminate/7 line 611
[06:52:27] <trentster> fxhp: hmm very odd indeed.
[06:52:35] <trentster> you got plenty of ram right?
[06:52:50] <trentster> I will test it out shortly.
[09:00:57] *** aortmann has joined #project-FiFo
[09:05:38] *** alcir has joined #project-FiFo
[09:19:51] *** mgls has joined #project-FiFo
[15:20:47] <fxhp> plenty of disk and ram
[15:46:51] <MerlinDMC> trentster, is that ghost thing a simple web app?
[15:58:49] *** jim80net has joined #project-FiFo
[16:03:28] *** Harvi has joined #project-FiFo
[16:09:00] <Harvi> Hello I have just installed 0.4.1 and after I start importing datasets the web gui starts giving me 500 errors. I have tried restarting nginx, epmd, snarl, sniffle, howl, wiggle also chunter on the hypervisor, im using b779b49a-29e4-11e3-9b1d-0b0b41ccdcad the web gui seems to be ok to browse the minute i start importing datasets it starts to http 500
[16:18:30] <MerlinDMC> Harvi, 500 for a request to what url? not something external like the dataset repository at datasets.at?
[16:21:27] <Harvi> internal api calls Error: 500 Object {method: "GET", url: "https://192.168.0.202/api/0.1.0/hypervisors"}
[16:21:52] <Harvi> pretty much on every menu link
[16:22:18] <Harvi> except the admin menu
[16:24:32] <Harvi> im currently working through http://project-fifo.net/display/PF/Problem+Checklist to see if i can nail the problem
[16:56:23] *** conan_the_destro has joined #project-FiFo
[16:58:12] *** aortmann has quit IRC
[17:23:45] *** tomww has quit IRC
[17:32:45] *** wiedi has quit IRC
[17:39:30] *** tomww has joined #project-FiFo
[17:39:41] *** tomww has joined #project-FiFo
[17:56:37] <killfill> hi
[17:57:13] *** alcir has quit IRC
[17:58:08] <killfill> ooOooo
[17:58:24] <killfill> instalation guide was updated to use 13.2.1...
[17:58:26] <killfill> !
[17:58:46] <killfill> Anyone knows how to migrate from a previous dataset to a new one?
[17:59:02] <killfill> just coping the db files should be enough?
[18:03:25] <killfill> Licenser: how was that thing one needed to make on the hypers? https://www.monosnap.com/image/rnHSgJpv8vqXMl7eFUts1pd04  :S
[18:03:37] <killfill> about the alias
[18:04:38] <killfill> echo something > somewher
[18:11:14] <killfill> oh got it..
[18:13:47] *** _lb_ has joined #project-FiFo
[18:21:53] *** fxhp has quit IRC
[19:23:54] <killfill> poor jingles.. doesnt have the fifo- prefix.. :P
[19:36:24] <killfill> oh yes it has
[19:46:04] <MerlinDMC> evening (here at least)
[19:55:13] <killfill> :)
[20:10:08] *** noahmehl has joined #project-FiFo
[20:47:40] <Licenser> hi guys
[20:51:57] <killfill> Hi Licenser, im about to re-define my network iprange on my new fifo, just like the old one. is is possible to manually set the next avail ip?.. so it doesnt conflict with my already created vm's
[20:52:20] <Licenser> killfill set the start ip?
[20:52:26] <killfill> yup
[20:53:56] <killfill> hm..
[21:33:12] <killfill> you be nice to have a 'dump' for the databases.. :P
[21:33:21] <killfill> `fifoadm dump`
[21:46:40] <MerlinDMC> my upstream at home sucks -.-
[21:46:54] <MerlinDMC> uploading 130M takes about an hour
[22:01:48] <killfill> anyone saw this crap? https://xen-orchestra.com/screenshots/
[22:04:45] <MerlinDMC> killfill, yep ... while I was seeking a frontend for our xen stuff at work ... back the it was not OSS tho
[22:05:55] <killfill> MerlinDMC: did you happend to use it?
[22:06:01] <MerlinDMC> but now that I'm migrating away from xen and do some SmartOS hosting i'm fine
[22:06:14] <MerlinDMC> no ... didn't want to buy smth :P
[22:06:40] <MerlinDMC> I did try archipel back then and started building up on that ...
[22:06:47] <killfill> i had to use xen for the office stuff (noboady is hosting smartos in USA aprarently)
[22:07:11] <MerlinDMC> archipel + ANSOS is quiet nice to deploy + use
[22:07:44] <killfill> and a guy show me that thing.. its completly unusable.. i mean.. leaving the smartos/fifo fanatic stuff behind.. its amazing.. and he told me xen was founding that gui
[22:08:09] <MerlinDMC> :)
[22:08:37] <MerlinDMC> it's always the marketing that gets you money ... not the actual project :P
[22:08:50] <killfill> true..
[22:09:11] <killfill> archipel looks greate
[22:09:33] <MerlinDMC> yeah ... I did like that quiet a lot
[22:09:52] <Licenser> fifo looks better :P
[22:09:54] <killfill> this 'competitors' should be in fifo's wiki.. :P
[22:10:01] <MerlinDMC> having all your machines in your jabber contact list and be able to issue commands in a chat was funny ^^
[22:10:28] <killfill> :)
[22:10:46] <killfill> would be cool as a holiday project, to put the fifo ruby api on hubot.. :P
[22:11:12] <Licenser> hubot?
[22:11:40] <killfill> http://hubot.github.com/
[22:16:15] <MerlinDMC> 17 minutes until the upload is done ... what to do now
[22:16:27] <MerlinDMC> have all movies watched :/
[22:26:37] <killfill> hm..
[22:27:00] <killfill> Licenser: fifoadm hypervisors list|grep r410a --->             r410a                               r410a
[22:27:10] <killfill> fifoadm hypervisors get r410a  ---> empty
[22:27:12] <killfill> :S
[22:27:28] <Licenser> killfill did you read the warning that you should not fill the file by hand?
[22:27:59] * killfill searching for the warning...
[22:28:32] <Licenser> it was in the mail :P
[22:28:41] <Licenser> basically: you should not do that
[22:30:01] <killfill> this one? https://groups.google.com/forum/?fromgroups#!topic/project-fifo/0vMFBxb5q7o
[22:30:08] <killfill> i actually did uname -n > /opt/chunter/etc/host_id
[22:33:03] <Licenser> just don't
[22:34:21] <Licenser> it will mostly just make things more complicated in the long run
[22:36:27] <killfill> ok.. should i just wipe that file, and delete all the hypers?
[22:38:37] <MerlinDMC> trentster, http://datasets.at/ui/#!/configure/9bc63934-3507-11e3-a7b4-fff9dc85302d ;)
[22:47:26] <Licenser> killfill it's the better approach
[22:57:34] *** _lb_ has quit IRC
[22:57:56] <killfill> hm...
[22:58:00] <killfill> Licenser: doesnt want to delete: fifoadm hypervisors delete r410a | fifoadm hypervisors list|grep 410a|wc -l
[22:58:03] <killfill> 1
[22:59:32] <killfill> am i doing something wrong?
[23:10:00] *** mgls has quit IRC
[23:12:21] <killfill> oh finnally can use the python client.. :)
[23:17:56] <killfill> Licenser: how do i wipe out thouse bad hypers? http://pastebin.ca/2466774
[23:25:31] <trentster> morning all.
[23:26:05] <trentster> MerlinDMC: http://ghost.org/features/
[23:27:17] <trentster> killfill: howdy, I noticed on latest dev, with nodes that have multiple zpool's no longer show hypervisor details when clicked on.
[23:29:36] <trentster> MerlinDMC: you are awesome dude!
[23:29:57] <killfill> trentster: oh.. wil not be ablt to replicate that..  gm...
[23:30:02] <trentster> :-D
[23:32:16] <trentster> killfill: https://www.monosnap.com/image/VpDzQUoOfQwPTDw86Ghy2govK
[23:48:01] <trentster> Harvi: did you definitely assign the full "…" permissions to your admin user when you setup?
[23:48:52] <killfill> Harvi is Diego?.. you speak spanish mate? :P
[23:49:08] <trentster> oops no :-P
[23:49:18] <killfill> trentster: uh :(
[23:49:28] <killfill> :P
[23:49:45] <trentster> maybe I should take some lessons
[23:50:04] <trentster> altho with the way the world is heading, Chinese may be more useful ;-)
top

   October 14, 2013  
< | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | >