Fell with AWS? Call in without question, documents later, now not before
While I was driving to work and listening to a new Dolphin album, someone blocked Amazon and Google’s IP addresses with entire subnets. Roskomnadzor called unreliable information about blocking sites that are not related to Telegram, but renting IP addresses on the same services as the instant messenger, and created a hotline to counter the spread of such messages.
Therefore, I can’t even guess why we had to help several services during yesterday’s day. Someone hurt quite a bit, and someone had to migrate.
In short, we can quickly migrate from AWS. We have a compatible API. Migrate customers from Google too. And in the coming weeks, we are ready to do the migration for free and give a month for tests in our cloud without a contract and without a letter of guarantee, simply by company card. Do not like it - do not pay anything.
')
Below - howto for different types of unloading VM from Amazon.
We are seeing a surge of interest in Russian clouds right now. We understand that those faced with blocking can either wait or urgently migrate their services to an accessible public cloud.
Migration requires access, arrangements with support, conclusion of a contract.
We are ready to skip these formalities and give access by letter to MBerezin@croc.ru.
The KROK cloud is developing largely under the AWS scenario: we ourselves develop our virtualization platform based on KVM. Naturally, in the process, we are learning from the experience of our Western colleagues, and therefore we can boast of compatibility with AWS, which we ourselves greatly respect.
Thesis:
Your customized automation will work in our cloud without any major problems. We have implemented the main functionality of the three services Elastic Compute Cloud (EC2), Simple Storage Service (S3), CloudWatch.
1. The standard procedure for importing / exporting a VM is described in the documentation .
However, there is an important limitation: exporting is possible only for machines that were previously imported into AWS. Otherwise, we get an error:
If you managed to export the instance, then you need to download it from AWS S3.
Next, download the downloaded image in the CROC S3. To do this, you need to download the CROC Cloud API access settings file, configure s3cmd to work with our S3 and load the disk image.
2. If you cannot export an instance (machines were created from AWS templates), then you can bypass the limitation as follows: the VM disk must be dumped into an image file, which will need to be loaded into our S3.
a) Create a temporary EBS volume slightly larger than the original one.
b) Create a file system on a temporary volume and mount it:
c) Install qemu-img
d) Stop the entire butt and services, remove the disk image:
e) Load the file /tmp/export/web1-xvda.qcow2 into CROC S3, create a template from it and start an instance from it in the same way as in the first version.
f) unmount and delete temporarily created EBS volume
3. Similarly for other OS, in Windows you can, for example, create VHD / VHDX through the disk control panel.
4. You can consider the option of backup and restore it on another machine.
5. If it is enough to transfer files (for example, content), rsync can be used in Linux, robocopy in Windows.
6. If downtime or large volumes are completely expensive, you can work with the replication of the OS and file systems through Double Take / Carbonite.
In general, there are quite a few options, it all depends on how the specific infrastructure is arranged. In all these cases, our support understands, and we have experience of similar migrations. Plus, our certified AWS specialists, so we can definitely help.
Have questions - please contact us in the comments or write: MBerezin@croc.ru
Therefore, I can’t even guess why we had to help several services during yesterday’s day. Someone hurt quite a bit, and someone had to migrate.
In short, we can quickly migrate from AWS. We have a compatible API. Migrate customers from Google too. And in the coming weeks, we are ready to do the migration for free and give a month for tests in our cloud without a contract and without a letter of guarantee, simply by company card. Do not like it - do not pay anything.
')
Below - howto for different types of unloading VM from Amazon.
We are seeing a surge of interest in Russian clouds right now. We understand that those faced with blocking can either wait or urgently migrate their services to an accessible public cloud.
Migration requires access, arrangements with support, conclusion of a contract.
We are ready to skip these formalities and give access by letter to MBerezin@croc.ru.
Compatibility
The KROK cloud is developing largely under the AWS scenario: we ourselves develop our virtualization platform based on KVM. Naturally, in the process, we are learning from the experience of our Western colleagues, and therefore we can boast of compatibility with AWS, which we ourselves greatly respect.
Thesis:
- Two access zones based on our own network of data centers,
- Equally fast network for all VMs, regardless of their type,
- Fast network between VM and physical infrastructure located in data center
- High performance fixed IOPS drives (analogous to io1 in AWS),
- Full service of import / export of virtual disks,
- Free user traffic in S3.
Your customized automation will work in our cloud without any major problems. We have implemented the main functionality of the three services Elastic Compute Cloud (EC2), Simple Storage Service (S3), CloudWatch.
Migration options
1. The standard procedure for importing / exporting a VM is described in the documentation .
However, there is an important limitation: exporting is possible only for machines that were previously imported into AWS. Otherwise, we get an error:
$ aws ec2 create-instance-export-task —instance-id i-XXXXXXXXXXXXXXXX —target-environment vmware —export-to-s3-task DiskImageFormat=vmdk,ContainerFormat=ova,S3Bucket=my-export-bucket An error occurred (NotExportable) when calling the CreateInstanceExportTask operation: Only imported instances can be exported. If you managed to export the instance, then you need to download it from AWS S3.
Next, download the downloaded image in the CROC S3. To do this, you need to download the CROC Cloud API access settings file, configure s3cmd to work with our S3 and load the disk image.
2. If you cannot export an instance (machines were created from AWS templates), then you can bypass the limitation as follows: the VM disk must be dumped into an image file, which will need to be loaded into our S3.
a) Create a temporary EBS volume slightly larger than the original one.
b) Create a file system on a temporary volume and mount it:
# mkfs.xfs /dev/xvdf1 # mkdir /tmp/export # mount /dev/xvdf1 /tmp/export c) Install qemu-img
d) Stop the entire butt and services, remove the disk image:
# qemu-img -O qcow2 -o compat=0.10 /dev/xvda /tmp/export/web1-xvda.qcow2 e) Load the file /tmp/export/web1-xvda.qcow2 into CROC S3, create a template from it and start an instance from it in the same way as in the first version.
f) unmount and delete temporarily created EBS volume
3. Similarly for other OS, in Windows you can, for example, create VHD / VHDX through the disk control panel.
4. You can consider the option of backup and restore it on another machine.
5. If it is enough to transfer files (for example, content), rsync can be used in Linux, robocopy in Windows.
6. If downtime or large volumes are completely expensive, you can work with the replication of the OS and file systems through Double Take / Carbonite.
In general, there are quite a few options, it all depends on how the specific infrastructure is arranged. In all these cases, our support understands, and we have experience of similar migrations. Plus, our certified AWS specialists, so we can definitely help.
Have questions - please contact us in the comments or write: MBerezin@croc.ru
Source: https://habr.com/ru/post/353704/
All Articles