最新の API linode.clone()を発表します。 この API コールを使うと、既存のLinodeを選択したデータセンターの新しいインスタンスに簡単にクローンできます。 クローン元のLinodeにあるすべての設定プロファイルとディスクイメージをクローンします。
このメソッドには、同じソースからのクローン同時作成は5つまでという制限があります。これは Linode インフラを保護するためです。同時実行の制限以上に別のクローン作成を実行しようとすると、API は"検証"エラーを返します。 大規模なクローンプロセスを実行したい場合は Linode クローンを 5 回作成し、それらのクローニングが終了すると、複製された Linode をそれぞれ 5 回同時に複製できます。 これによりインフラを単一の Linode から数百にすばやくスケールすることができます。
この新しい API コールを活用し、これが Linode クラウドの将来の展開に役立つことを願っています。
- Tim
コメント (10)
Wow, great!
We’ve been looking for something like this for a while
Are there any pre-requisites for this API Call?
Is the Linode need to be shutdown or will be shutdown during the clone process?
How long would it take to clone? Is it the same thing as if its done from the console?
AWESOME! Been waiting for this for so long, at least now we can script our scaling activities!
Can we get one for checking the amount of bandwidth/transfer used for the current month next? (To compare against the existing call for finding the quota.)
@Dash
> Is the Linode need to be shutdown or will be shutdown during the clone process?
https://www.linode.com/api/linode/linode.clone —
“It is recommended that the source Linode be powered down during the clone.”
The source Linode doesn’t have to be powered down, but it is recommended to ensure that a consistent copy of your data is cloned.
> How long would it take to clone? Is it the same thing as if its done from the console?
1-2 minutes per GB of data within the same datacenter of the source Linode (the Linode you’re cloning from), or 5-10 minutes per GB of data if migrating to a different datacenter.
I assume that you should not .clone() a Linode that has static networking set up, since the static confi will render the new node unreachable… or am I missing something?
Xof:
You could still clone the Linode, but you’d need to use LISH to log in to the new Linode and adjust the networking configuration to the new IP address(es).
Tony Yarusso:
There isn’t currently a way to see how much transfer a particular Linode has used that I can see, but account.info() shows transfer pool amount and usage:
https://www.linode.com/api/utility/account.info
How would you suggest auto scaling in an HA environment where you want to clone and configure network addresses all in one go?
The limitation of 5 clones per source linode seems very arbitrary – you’re just offloading additional work to the side of the script doing the cloning. You might think we’re going to keep track of the clone count on our own, but the truth is we’re just going to iterate over existing linodes until the clone() call returns without an error.
What difference could it possibly make on your end whether we clone a single linode 10 times versus cloning an initial linode 5 times, and then cloning the first clone 5 additional times? There should be no difference in your ability to “protect Linode infrastructure”..?
Xof brought up a good question regarding the static networking setup causing new node unreachable. May I know what is the best solution for that in Linode? Doug’s suggestion to login using LISH and change it manually doesn’t make sense to me, as it defeats the purpose of autoscaling.