Been trying out mercury & all seems to be working quite well but now im on the snapshots bit & have a question... I followed your instructions for a single ebs drive http://groups.drupal.org/node/33092 and it all seemed to work except the consistent snapshots stuff... What is the replacement pattern for the command you listed here http://groups.drupal.org/node/36766 ... ive tried adding the relevant bits in that command but keep getting the following error.
ec2-consistent-snapshot: ERROR: 500 Can't connect to us-east-1a.ec2.amazonaws.com:80 (Bad hostname 'us-east-1a.ec2.amazonaws.com')This is my first mercury setup & firstish ec2 instance.... so im thinking either im missing some important piece of the puzzle or I am typing out the command from your post above incorrectly, or both lol; which parts of the command do I change? the capitalized parts?
My mercury instance is standard ive got a volume and an elastic ip attached to the mercury ami, pressflow site is accessible and ssh & mysql stuff moved to /vol afaict.
Also I was wondering if you could you share an example of a cron command with this goody in it?
Cheers!
Comments
my best guess is that your
my best guess is that your region is off somewhere.
you have to replace the
--region us-east-1part with the region that your instance and EBS are in for the command to work. To the best of my knowledge this doesn't require an "a" "b" "c" or "d", just the first part of the region and the #.
here's the crontab I'm using
crontab -e
0 0,12 * * * <command here>
I just created a text file in my user directory added the command to it, made it executable (chmod +x ) and then added it with the above crontab command.
That crontab runs the command to snap shot my system once every 12 hours (at 0:00 and 12:00) so I get two snapshots per day.
I'm going to work on a maintenance bit to add to the script that deletes snap shots older than X days when it runs so I don't have to go in and clean them out so often but I've not had the time to get that done yet.
Doug
@nullvariable | www.nullvariable.com
I just updated the
I just updated the instructions here: http://groups.drupal.org/node/36766 and added more detailed instructions on creating a cron job to the wiki page.
Doug
@nullvariable | www.nullvariable.com
Sweet!
Thanks for the more detailed instructions... im pretty sure I mucked up the command and edited the wrong spots, I did have the region right though. I will give it a whirl as soon as my office is set up again.... been moving the last couple of days.
Thanks for sharing your method on the crontab that's a nice way to handle it, ive not used crontab that way before.
Im totally loving AWS this shit is cool! ...and my initial tests of the pantheon site were blazing fast, I cant wait to dig in it more to see how its all functioning. Im excited to get my Aegir and sites into pantheon and a load balancer setup to carry me over for the next several months until I split of some of the main server stuff into their own instances.
I love having this whole ami with all the bits working to study. Its especially nice to be able to dig into solr, we wanted to use this for our real estate site were soon to launch, and it helps a ton to see a functioning example. To able to so easily drop our dev site into pantheon and test & tweak it from a great starting place is awesome! You guys rock so much for releasing this!
Cheers!
just ran into the same thing
just ran into the same thing now that I have a raid setup and was going for the snapshot, you were right it was the region and I didnt need to include the ending letter.
Are there any considerations when restoring? I looked in 'man ec2-consistent-snapshot' and there was no mention of it. Do you have to re-create the whole setup again from the snapshots? Create new drives off the snaps then attach and then reconfigure the raid with mdadm? That sounds about right AFAICT... too bad they dont have a method for easy bouncing between snapshots... or do they?
I'm loving that this uses differential backups and only snaps whatever has changed since the last snap totally sweet!
Thank you so much for your tutorial its really helped me jump in and get my hands dirty... and all seems to be working quite well.
yes. I am planning to work
yes. I am planning to work through the restore process but I've been too busy to do so yet. Hopefully I'll get it all worked out before I have to do it because a machine went down!
My assumption is that in the future I'll have it scripted so that cloudwatch can boot up a new instance and run a script for me that will automatically load this configuration onto a new instance and mount the drives. Snapshots would only really be needed if I wanted to duplicate the instance for dev or testing work, or if the site was hacked and I needed to go backwards in time. Drive failure is supposed to be very unlikely compared with EC2 instance failure.
Doug
@nullvariable | www.nullvariable.com
Yeah cloudwatch looks quite
Yeah cloudwatch looks quite interesting ... especially the auto scaling bits & ability to auto-run scripts to handle all manner of things; that and the elastic load balancer are yet to be implemented in my setup, im holding off on it until next year when we will really need it.
The built in ebs redundancy with multiple data replication and all is fantastic, it makes the raid0 setup viable & seems to be faster than using the ephemeral store & certainly safer. I moved aegir, my web-root, ami tweaks and the database stuff in and feel much better getting those snapshots, makes me sleep a lot better that's for sure!
this php script is really
this php script is really handy when you end up with 275 snapshots after regular backups :)
http://code.google.com/p/ec2-delete-old-snapshots/
Doug
@nullvariable | www.nullvariable.com
Sweet ... thanks for the link
Sweet ... thanks for the link .. I will try that for sure, ive been doing it by hand every so often .... which can be time consuming!
Cheers!