And sudo docker-compose up -d
to start it up again
Yeah, you can just do sudo docker-compose down
to shut down the server
ComfortableHorse5 just to make sure, I just spun up an AWS instance with the ClearML Server 1.0.1 AMI Community AMI, and:
` > cd /opt/clearml
ll
total 0
drwxr-xr-x 8 root root 153 May 14 09:46 agent
drwxr-xr-x 2 root root 6 May 11 10:21 config
drwxr-xr-x 6 root root 67 May 11 10:21 data
drwxr-xr-x 2 root root 49 May 11 10:22 logsSo I'm not sure why you didn't see an
/opt/cleaml/configdirectory... Also, the
docker-compose.ymlfile is located in the home direcotry:
> cd ~
ll
total 8
-rw-r--r-- 1 ec2-user ec2-user 4646 May 11 10:21 docker-compose.yml
`
The AMI is simply a pre-built instance, with the docker compose deployed in it - you're probably not looking in the right place.. 🙂
Hi ComfortableHorse5 ,
No, unless you need to override any settings, these default settings are already part of the built-in server.
I’ve launch the ami a few times however, it seems to blank.
Can you provide more info? What do you see when you surf to http://<server-public-address>:8080
? Note it might take a few mins for the server to finish booting up (especially due to the load time of ElasticSearch)
yes UI is normal i can browse around without problem
OH, I see 🙂
Well, by default (as you can see in the docker.compose.yml
file, in the services.apiserver.volumes
section), the server will look for extra configuration files in the host's file system under /opt/clearml/config
. If it's not there, no harm is done.
You can simply shut down the server, create the /opt/cleaml/config
folder (make sure it has the same permissions as the rest of the /opt/clearml
folder where the server is currently deployed), add the required apiserver.conf
file, and start the server up again.
By the way, which server version are you using?
seems like i had to restart the entire instance instead
sudo docker-compose down sudo docker-compose up -d
don’t think it works that way with this AWS EC2 AMI image, there is no docker.compose.yml file to begin with.
oh yeah, you’re right, i didn’t even check the home directory … my bad..
ic, when you say shutting down the server, you’re refering to the dockersclearml-webserver clearml-apiserver clearml-fileserver
?
I’ll just stop them and create those *.conf files accordingly and start the dockers again ?
OK, and than? Up to this point this is the normal behaviour
sorry for not being clear. Let me detailed that as much as i can:
- Refering to https://allegro.ai/clearml/docs/docs/deploying_clearml/clearml_server_aws_ec2_ami.html . I launch my own EC2 using this “ap-northeast-2 : ami-049f5d9fc4f64d0a8” as my AMI
- After EC2 is launched, browsing to it directly via IPADDRESS:8080 is alright for my end.
- I then wanted to customize/configure the ClearML Server to use sub-domains instead of ip address. Hence, i’m following this instructions here https://allegro.ai/clearml/docs/docs/deploying_clearml/clearml_server_config.html?highlight=domain#sub-domains-and-load-balancers
- With that, i SSH into the ClearML Server box, and browsed to /opt/clearml/config
- However, i didn’t find any files listed in /opt/clearml/config
- I thought, somethign might be wrong in how i launch the AMI or configuration. Therefore, i terminated the ClearML Server, and try again repeating step1.
- Tried a few times, and waiting for a while before i SSH again into the box to check again, however browsing to /opt/clearml/config still didn’t give me any listing of apiserver.conf, and the other expected *.conf files.
But of course you're right in that restarting the instance is much easier and an easy solution 😄
So what did you mean by "I’ve launch the ami a few times however, it seems to blank"?
http://server-ip:8080 browse me into ClearML login page asking me for my name.