[00:20:57] *** benbangert has joined #project-FiFo
[00:21:07] <benbangert> I don't suppose anyone has gotten the community gitlab image working?
[00:34:19] <fxhp> benbangert - I wrote a salt-state for rhodecode (pylons based)
[00:34:31] <benbangert> cool
[00:34:52] <fxhp> (not a zone though, its a KVM)
[01:37:14] *** ira has joined #project-FiFo
[01:39:11] *** trentster has joined #project-FiFo
[01:42:29] <trentster> benbangert: I am using the community gitlab image. I also had some weirdness with it innitially, but can tell you what needs to be done to get it sorted ;-)
[01:42:44] <benbangert> I figured it out, I hadn't allocated enough ram
[01:42:47] <benbangert> it didn't like 1GB
[01:42:54] <benbangert> using 2GB, it now seems to be running ok
[01:43:09] <benbangert> email doesn't work, but that seems related to google being picky about what connections it allows
[01:43:34] <trentster> yeah, it needs more ram, and also I had to increase the timeout for the ruby compile tasks on innitial startup.
[01:43:49] <benbangert> ah yea, first time I hit it, it failed
[01:43:52] <benbangert> took 31 sec, heh
[01:43:56] <benbangert> but after that its been fine
[01:43:59] <benbangert> maybe my machine is faster
[01:44:01] <trentster> heh, murphys law huh?
[01:44:59] <trentster> MerlinDMC: is going to redo the image with more sane timeouts by default, as I am sure a lot of people run into this, and just figure the dataset is busted somehow and move on.
[01:45:28] <benbangert> maybe have the description mention the min ram it should be allocated
[01:47:10] <trentster> yeah, hopefully in newer version of datasets.at, there will be support for including something like an optional "MD" file with a dataset upload that contains dataset specific pertinent info.
[01:47:22] <trentster> I think it would be a good addition.
[01:52:19] <benbangert> definitely
[01:55:45] <benbangert> MerlinDMC: nice job on the gitlab dataset, I dig it :)
[01:56:38] <trentster> yeah its awesome
[01:56:50] <trentster> and he is sleeping atm ;-)
[03:13:13] *** ira has quit IRC
[04:22:00] *** fxhp has quit IRC
[04:37:13] *** noahmehl has joined #project-FiFo
[05:18:54] *** adhioutlined1 has joined #project-FiFo
[05:19:01] <adhioutlined1> hi all
[05:19:25] <adhioutlined1> is it possible to add IP range in fifo without GAteway defined ?
[05:20:34] <adhioutlined1> I mean add a new IP range in jingles
[05:25:11] *** MerlinDMC has quit IRC
[05:34:38] *** MerlinDMC has joined #project-FiFo
[07:30:20] <MerlinDMC> trentster + benbangert, the gitlab dataset needs 2G ram ... at least for the first requests to compile all the stuff (datasets.at does set a default of 2G ram in the UI) - the timeout thing will be fixed once joyent does release the new base images and maybe then also gitlab 6.2 is released
[07:32:06] <MerlinDMC> morning btw
[07:32:13] <benbangert> or evening :)
[07:32:27] <benbangert> yea, I noticed that on 2G, once I did that it all worked swimingly well
[07:32:32] <MerlinDMC> yeah ... you guys at other sides of the world - can't do much about that ;)
[07:32:43] <benbangert> except for email sending, but I think its my ISP nuking outbound port 25 connections
[07:32:53] <MerlinDMC> email works fine here
[07:33:01] <benbangert> gonna have to set it up to relay through a google account of mine
[07:33:06] <benbangert> yea, I'm sure its my ISP
[07:33:30] <MerlinDMC> it does use the local postfix ... so if you just configure that for a relay you should be fine
[07:33:44] <benbangert> yea
[07:33:50] <benbangert> I found some instructions on that
[07:33:54] <MerlinDMC> if you did change your stuff there ... you could tell me the changes and i'll integrate some metadata setting for a mail relay
[07:34:32] <MerlinDMC> the dataset build is done via mibe ... so it'S all templatized
[07:37:01] <MerlinDMC> ... I remember your name from somewhere ... freaking me out that I can't remember
[07:38:29] <benbangert> lol
[07:38:41] <trentster> MerlinDMC: howdy mate.
[07:38:50] <benbangert> MerlinDMC: I've been in here before?
[07:39:04] <benbangert> I think I was mentioning that the datasets were missing ones joyent added or something
[07:39:07] <MerlinDMC> yeah ... that I did realize ^^
[07:39:15] <trentster> benbangert: email works for me as well, I just added smart relay to postfix main.cf and all works perfecto
[07:39:27] <MerlinDMC> oh yeah ... that it was
[07:39:30] <benbangert> and the graphs weren't working in project-fifo, which the latest project-fifo update actually fixed. graphs now work :)
[07:40:32] <trentster> benbangert: hmm, when you say graphs you don't mean cloudview? you eman like the cpu and nic stats for vm's right?
[07:40:41] <benbangert> yea
[07:40:46] <MerlinDMC> trentster, does the ikvm client on the supermicro ipmi stuff work for you? ... I can't attach an iso while on my mac - works fine on windows tho
[07:40:51] <benbangert> the cpu/nic stats for vm's and such. none of the graphs worked before
[07:41:02] <benbangert> latest update fixed them, now they all update/display data
[07:42:01] <trentster> MerlinDMC: I havent tried to be honest
[07:42:20] <benbangert> tempted now to get more RAM, fifo has made it so easy to spin up more VM's I'm starting to hit my limit. hehe
[07:42:23] <trentster> benbangert: yeah worked for me before and they work for me now;-)
[07:42:40] <MerlinDMC> would you do for me quickly? ^^ ... I can select the virtual cdrom tab ... change the selectbox to iso and if I click open image it just gets stuck
[07:42:42] <trentster> However that being said, cloudview is now broken, I think it may be removed going forward
[07:42:52] <trentster> not sure what killfill thoughts on that are?
[07:43:08] <benbangert> cloudview for me is..... oriented oddly
[07:43:18] <benbangert> its a small clipped window in the top left
[07:43:26] <trentster> benbangert: real limit or fifo imposed limit including powered off machines?
[07:43:42] <benbangert> oh.... dunno
[07:43:52] <benbangert> it counts powered off ones?
[07:43:59] <MerlinDMC> yeah
[07:43:59] <trentster> benbangert: yeah me to, I moved 3 fifo instances to 13.2.1 datasets today, and its doing that for me as well on all 3 of them.
[07:44:04] <MerlinDMC> because the ram is provisioned
[07:44:35] <MerlinDMC> I don't like that behaviour because I have a lot of dataset templates around :P
[07:44:55] <benbangert> wait, the ram is marked as unusable to smartos for non-running zones?
[07:45:11] <benbangert> or is this just PF that sees it as used?
[07:45:31] <benbangert> oh, just PF
[07:45:34] <MerlinDMC> not to smartos ... vmadm doesn't care in the newer versions
[07:45:47] <benbangert> yea, I'm running this for my home server :)
[07:45:54] <benbangert> my 'cloud' is a machine of one. hehe
[07:46:11] <trentster> soem nice humour there from Licenser
[07:46:18] <benbangert> I spin up lots of little VM's that have different apps in them as one-off's for various dev testing stuff
[07:47:56] <benbangert> hmm, still getting close even when I count the ones that are off though
[07:48:29] <benbangert> still got about 5GB of ram free says PF though, so that should be fine for awhile
[07:49:45] <MerlinDMC> benbangert, and you did add 1G for every kvm?
[07:50:00] <benbangert> what do you mean?
[07:50:20] <MerlinDMC> a kvm zone where you set 1G ram reserves 2G
[07:50:31] <MerlinDMC> 1G for qemu and 1G for some caches inside the zone
[07:51:33] <benbangert> oh
[07:51:53] <benbangert> lesse, one KVM has 512MB, and one KVM has 2GB
[07:51:57] <MerlinDMC> that second "cache" 1G is not locked down tho ... it can be shared but fifo counts it
[07:51:58] <benbangert> the rest are zones
[07:52:30] <benbangert> neither of the KVM's were provisioned or setup with fifo
[08:51:29] *** adhioutlined1 has quit IRC
[08:57:43] *** adhioutlined has joined #project-FiFo
[09:24:05] *** alcir has joined #project-FiFo
[09:36:09] <Licenser> MerlinDMC you're sure that it can be shared? as in can I run two 512 KVM machines with 1.5G of memory?
[09:42:59] *** mgls has joined #project-FiFo
[09:48:08] <MerlinDMC> Licenser, yes ... it's just undefined what happens if both zones try to use up 100% of that additional 1G ... but that should never happen
[09:48:36] <MerlinDMC> same as if you boot smartos on a 4G server and create a 128G zone ... works
[09:52:20] <Licenser> hmm did that change? I remember that you could not start VM's / create Vms if you had too little memory
[09:52:46] <Licenser> even when ZFS was using it for cache, giving something 'not enough memory error'
[09:54:19] <MerlinDMC> they changed the logic - they removed it so the actual user is responsible
[09:54:52] <MerlinDMC> but was possible before as well through vmadm ... maybe they had some intermediate version where it was blocked - don't know
[11:37:16] *** adhioutlined has left #project-FiFo
[14:27:49] *** noahmehl has quit IRC
[15:16:15] *** jim80net has joined #project-FiFo
[17:25:08] *** alcir has quit IRC
[19:16:50] <killfill> Hey Licenser, why does howl use riak_cluster?
[19:17:08] <killfill> is it for saving the states of the subscribed channels?
[19:17:32] <killfill> so one can possible scale up?
[19:17:34] <Licenser> riak_core not riak_cluster but for two reasons: 1) recover from failure of a single node 2) loadballancing communications
[19:20:11] <killfill> Ag
[19:20:13] <killfill> Ah
[19:20:50] <killfill> Thought only sniffle and snarl did.. :P
[19:21:22] * killfill should learn riak_core sometime..
[19:25:39] <MerlinDMC> huh ... osx mavericks got even cheaper ... funny
[20:34:40] *** xxxlalala has joined #project-FiFo
[20:46:25] <xxxlalala> congrats on the interface it looks realy nice
[21:38:58] <Licenser> xxxlalala thanks killfill will love to hear taht
[21:39:40] <Licenser> MerlinDMC for osX 11.0 they'll give you 50 euro to download it ;)
[21:49:21] *** MerlinDMC has quit IRC
[21:49:33] *** conan_the_destro has quit IRC
[21:50:16] *** conan_the_destro has joined #project-FiFo
[21:51:02] *** MerlinDMC has joined #project-FiFo
[23:09:19] *** xxxlalala has quit IRC
[23:10:02] *** mgls has quit IRC
[23:22:01] *** noahmehl has joined #project-FiFo