flaccid | heh insecure | 03:41 |
---|---|---|
superxgl | hi all, i forgot to terminate the running VMs before stop the cloud. and when i start the cloud again , "euca-describe-availability-zones verbose" shows " free is 0000" so that i do not have enough resource to start vm , what should i do ? | 14:19 |
superxgl | oh, now i can start vm again, | 14:27 |
superxgl | don't know what happened.. | 14:28 |
=== daker_ is now known as daker | ||
kjdash | already here | 20:22 |
kim0 | kjdash: Hi | 20:22 |
kim0 | kjdash: you mentioned you want to work on the ebs migration tool right | 20:22 |
kjdash | yes | 20:22 |
kim0 | great .. so what are your questions | 20:22 |
kjdash | so the basic idea is: start an ebs instance in one region, and then start another in another region, copy the ebs volume over, snapshot and register | 20:24 |
kim0 | yes | 20:24 |
kjdash | do we require the user to list the amis to be launched in each region? | 20:25 |
kim0 | the ebs vol you copy over, is however not the ebs vol that is the root for the running instance | 20:25 |
kim0 | so both instances are running .. and each of them, has a ebs volume attached to it | 20:25 |
flaccid | use rsync | 20:25 |
kjdash | an ebs volume other than the root ebs volume | 20:25 |
kjdash | flaccid, yes, most likely | 20:26 |
kim0 | yes, i.e. the version of the utility ec2 instance, is independent than the the version of the ebs instance you're actually copying | 20:26 |
kim0 | i.e. you could launch 2 ubuntu instances to copy a centos instance, makes sense ? | 20:26 |
kjdash | so really, the goal is to migrate an ebs snapshot, not an ami | 20:27 |
kim0 | An ebs ami is an ebs snapshot (someone correct me if I'm wrong) | 20:27 |
flaccid | yeah just rsync. you can't get the snapshot out of s3 manually | 20:28 |
kim0 | flaccid: do you have the perfect rsync command for that ? copying all the special files and magic ? | 20:30 |
kjdash | ok, so here is where i see the problem | 20:30 |
flaccid | rsync -az does the trick depending on what you need to do | 20:30 |
kjdash | rsync -azv or some such | 20:30 |
flaccid | there is an article i have to review today on this topic that i can show you later on, though alestic.com and google.com has much on this | 20:31 |
kjdash | if you want to register an ami in the other region, you need to know things like kernel, ramdisk, architecture | 20:31 |
kim0 | copy from src ? | 20:32 |
kjdash | which is no problem if you require that information at runtime | 20:32 |
flaccid | personally, i don't re-use the root ebs volume as i design appliances | 20:32 |
kjdash | i would imagine something like the syntax of ec2-register-img but with --source-region and --dest-region options | 20:34 |
flaccid | the --region param. its in teh commandline reference doc | 20:34 |
kjdash | flaccid: i typically use a modified base image, and then a custom repository with meta packages | 20:35 |
kjdash | what do you use? | 20:35 |
flaccid | RightScale ServerTemplates | 20:37 |
kim0 | smoser: someone working on the ebs migration tool, asks if it's ok to target python 2.7 (for argparse module). Any ideas | 20:57 |
kjdash | so euca2ools doesn't support registering an EBS ami eh? | 21:46 |
flaccid | ec2-register | 21:47 |
flaccid | err euca-register | 21:47 |
kjdash | sure, but euca-register doesn't support snapshots | 21:48 |
flaccid | check the docs and see what does; sorry i don't use euca much yet | 21:50 |
kjdash | i was more asking for clarification | 21:50 |
kjdash | i looked at the source, it clearly does not support registering ami with snapshot | 21:51 |
flaccid | kjdash: http://www.google.com.au/search?client=opera&rls=en&q=eucalyptus+register+ebs+ami&sourceid=opera&ie=utf-8&oe=utf-8&channel=suggest | 21:51 |
flaccid | the first result explains | 21:52 |
kjdash | unf, the 6 searches i tried failed to produce that result. thanks | 22:05 |
flaccid | np | 22:07 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!