[01:02] <smoser> patrickw, no way to set it on per-instance
[01:03] <smoser> i believe it is configurable for a give size
[12:54] <Makere> how do I know if the node have registered correctly and the cluster is using them?
[14:07] <TeTeT> Makere: try euca-describe-availability-zones verbose and sudo euca_conf --list-nodes
[15:33] <TeTeT> hggdh: just a quick update, I got init scripts for apache and wget, a broker for recording and querying servers, and a scheduler written. now I'm trying to integrate these. hopefully done by end of tomorrow
[15:34] <hggdh> TeTeT: fantastic :-)
[15:39] <TeTeT> hggdh: do you know of a reliable way to get eth0 ip address from python? I use ifconfig eth0 | grep inet\ addr | sed \"s/:/ /g\" | awk '{ print $3 }' but it fails
[15:40] <hggdh> TeTeT: not really, no
[15:41] <hggdh> but
[15:41] <hggdh> hold on
[15:42] <hggdh> TeTeT: this seems to work: ifconfig wlan0 | grep "inet addr" | sed "s/:/ /g"|awk '{print $3}'
[15:44] <hggdh> as this: ifconfig wlan0 | grep "inet addr" | tr ":" " " | awk '{print $3}'
[15:54] <Makere> what do you use clouds for btw? Doing a school thing and we haven't really decided yet what to try/use 60-70 PC cloud with
[15:54] <TeTeT> hggdh: I found a python function that does it, will use that
[15:54] <TeTeT> hggdh: I encounter this problem fairly often, like 1 out of 10 launches with -f <script>: http://pastebin.ubuntu.com/525732/
[15:55] <TeTeT> Makere: I only use it for testing UEC and giving the UEC classes, so only lab work, not production
[15:55] <Makere> so in the end the whole thing is pretty much useless
[15:55] <Makere> ;P
[15:56] <hggdh> Makere: well, not really, we still find errors ;-)
[15:56] <hggdh> TeTeT: this is on 1.6.2, or 2.0?
[15:56] <TeTeT> hggdh: maverick, 2.0
[15:57] <Makere> is it possible to run one VM with all the nodes CPU's/Mem/HDD compiled? some sort of cluster
[15:57] <TeTeT> hggdh: I updated the training cloud to it, so it was lucid first, then upgrade to maverick. Upgrade went smoothly
[15:57] <hggdh> TeTeT: at least that... can you open a bug against cloud-init on this?
[15:58] <hggdh> Makere: you mean run a cloud on a VM?
[15:58] <TeTeT> hggdh: directly upstream at cloud-init on LP? Can do that
[15:58] <Makere> I mean forming a cluster with cloud
[15:59] <hggdh> TeTeT: at LP first; if no dups, then of course we would be happy to also have an upstream bug
[15:59] <hggdh> Makere: it should be possible, yes
[16:00] <Makere> 2 things I hate with cloud is, one that everything should be possible but in reality it's probably really annoying shit, second that all the material related to clouds seem to be like 90% bullshit
[16:02] <TeTeT> hggdh: bug 671000 filed against ubuntu/cloud-init
[16:02] <hggdh> Makere: mind your language, please.
[16:03] <hggdh> Makere: (1) not everything is possible; (2) there is a lot of hype around clouds, this is for certain
[16:04] <hggdh> Makere: for example: can you set a cluster in a cloud (which is different from clustering the cloud itself)? Yes. But I have never done it
[16:05] <hggdh> theoretically it should be possible, it is just a question of setting up, which should be pretty similar to a cluster on bare-metal
[16:06] <hggdh> TeTeT: can you please add versions of cloud-init and euca?
[16:09] <TeTeT> hggdh: ok, as soon as I get an instance up again
[16:13] <TeTeT> hggdh: hmm, I was using a lucid instance, maybe upgrade to maverick there?
[16:15] <hggdh> TeTeT: you can try with a maverick one, but a lucid image should also work
[16:15] <hggdh> if it does not, this is good data for the bug
[16:33] <TeTeT> hggdh: I pushed the scripts here: https://code.launchpad.net/~tspindler/junk/uec-scheduler
[16:34] <TeTeT> hggdh: I'm testing them now, hope the integration of them will be smooth
[16:34] <hggdh> TeTeT: cool, thank you, will branch from there
[16:44] <smoser> TeTeT, i believe your bug is a dupe of bug 571271 which was fix-released in 0.5.10-0ubuntu1.1
[16:46] <smoser> TeTeT, also, regarding https://bugs.launchpad.net/ubuntu/+source/cloud-init/+bug/671000
[16:47] <smoser> it is marked as "Maverick", but the version reported (0.5.10-0ubuntu1) is lucid GA (current in lucid-updates is 0.5.10-0ubuntu1.3)
[16:48] <smoser> oh. i see, maverick host.
[16:49] <TeTeT> smoser: I'm using a maverick image now and it seems to work just fine
[16:50] <smoser> right, as will more current lucid images.
[16:50] <smoser> i'm guessing the one you have is from the imagestore, right?
[16:50] <TeTeT> smoser: yes, that's what I show the students during class usually
[16:50] <smoser> which brings us to the next question of "Why are there not updated images in the store"
[16:50] <smoser> ...
[16:50] <smoser> which i'm hoping nijaba is going to address sometime shortly
[16:50] <TeTeT> smoser: the image store is a bit useless right now, IMO
[16:51] <smoser> TeTeT, useless due to out of date images ?
[16:51] <smoser> or other?
[16:51] <smoser> nijaba, has shown an interest in getting the out-of-datedness addressed.
[16:51] <TeTeT> smoser: yes, due to out dated images. I also wish I could delete an image from the store, so I can demo the installation to students more often
[16:53] <smoser> TeTeT, "delete an image" meaning remove the data that says this image has been installed
[16:53] <smoser> right ?
[16:53] <TeTeT> smoser: and the image from the bucket
[16:53] <smoser> i'm guessing theres some data on the CC that you could wipe to do that. the image from the bucket wouldn't really hurt, and can be deleted other ways, right ?
[16:53] <TeTeT> smoser: I can manually delete it from the command line, but afaik it is still displayed in the shop
[16:54] <TeTeT> smoser: what happens if you d/l the image again? Will the old data just be overwritten?
[16:55] <smoser> yeah, i'm guessing you can kill some data
[16:56] <smoser> you might try:
[16:56] <smoser> sudo apt-get --purge remove python-image-store-proxy  && sudo rm -Rf /var/lib/image-store-proxy/ && sudo apt-get install python-image-store-proxy
[16:56] <TeTeT> smoser: I don't have root privileges on the main training machines for the students
[16:57] <TeTeT> smoser: just eucalyptus admin rights
[16:57] <smoser> hm.. yeah, then from that you proably can't do it.
[16:57] <TeTeT> which I believe is a valid use case, that UEC admin and root are different entities
[17:04] <smoser> TeTeT, at very least it would make sense to be able to kick the image store to tell it to check its inventory
[17:04] <smoser> TeTeT, you could open a bug :)
[17:05] <TeTeT> smoser: I feel that I opened to many bugs already
[17:11] <TeTeT> hggdh: the scripts seem to be running, I check if this is still the case tomorrow morning. out now
[17:13] <smoser> hm.. TeTeT, i don't think anyone feels you've opened too many bugs.
[17:13] <smoser> (i realize you're gone)