Incompatible version between client and server

[imrefazekas]

I'm using the v3.6.0 builds of Nebula and the tool gives me this when I try to back up:

{"addr":{"host":"127.0.0.1","port":62301},"error":"Incorrect frame size (1213486160)","level":"error","msg":"Incompatible version between client and server.","time":"2023-09-19T16:05:42.764Z"}

Any ideas how to make it work?
Or which version I should move to?
Thanks!

[QingZ11]

Use the latest release v3.5.0 version of br.

[imrefazekas]

I'm using 3.5 now.
Built.
Agent is running as screenshot shows.

Yet the br is saying:

bin/br backup full --meta "127.0.0.1:49619" --storage "local:///Users/.../nebula/backup/"
{"level":"info","meta address":"127.0.0.1:49619","msg":"Try to connect meta service.","time":"2023-11-06T22:29:53.716Z"}
{"level":"info","meta address":"127.0.0.1:49619","msg":"Connect meta server successfully.","time":"2023-11-06T22:29:53.721Z"}
Error: parse cluster response failed: response is not successful, code is E_LIST_CLUSTER_NO_AGENT_FAILURE

Any suggestion?

[kqzh]

I'm using 3.5 now. Built. Agent is running as screenshot shows.

Yet the br is saying:

bin/br backup full --meta "127.0.0.1:49619" --storage "local:///Users/.../nebula/backup/"
{"level":"info","meta address":"127.0.0.1:49619","msg":"Try to connect meta service.","time":"2023-11-06T22:29:53.716Z"}
{"level":"info","meta address":"127.0.0.1:49619","msg":"Connect meta server successfully.","time":"2023-11-06T22:29:53.721Z"}
Error: parse cluster response failed: response is not successful, code is E_LIST_CLUSTER_NO_AGENT_FAILURE

Any suggestion?

hi, each service node needs to deploy an agent, can you check it?

[imrefazekas]

I tried to run another 2 agents connecting to different meta servers.

I can still see just one in Nebula Console web app...
And the br still says the same thing...

[kqzh]

I tried to run another 2 agents connecting to different meta servers.

I can still see just one in Nebula Console web app... And the br still says the same thing...

If you have multiple Metad or Agent services, you'd better use a real address instead of 127.0.0.1, you can execute SHOW HOSTS META to check meta addr

[imrefazekas]

That meta address shows the name of the host inside the docker and I want to back up from outside the docker

[imrefazekas]

hosts modified. agent is running

./agent --agent="127.0.0.1:8888" --meta="metad0:50543"

seen in the nebula studio:

and hosts in nebula studio:

and the br still tells:

parse cluster response failed: response is not successful, code is E_LIST_CLUSTER_NO_AGENT_FAILURE

[wey-gu]

Sorry, I know this is against common sense, the checking criteria is expecting every host comes with an agent, including the stateless graphd. I had been there, too for the first go.

https://github.com/wey-gu/nebula-up/blob/main/backup_restore/docker-compose.yaml

@kqzh could we consider removing the requirement on graphd hosts to come with an agent, too?

[kqzh]

Sorry, I know this is against common sense, the checking criteria is expecting every host comes with an agent, including the stateless graphd. I had been there, too for the first go.

https://github.com/wey-gu/nebula-up/blob/main/backup_restore/docker-compose.yaml

@kqzh could we consider removing the requirement on graphd hosts to come with an agent, too?

hi wey and @imrefazekas , during the recovery process, the agent will start and stop the graphd, therefore, an agent is needed on the host of graphd.

[imrefazekas]

@wey-gu thanks for the link. I used the docker compose linked about (little customisation though).

sleep 10 &&
        mc alias set my_object_storage http://nginx:9000 minioadmin minioadmin &&
        mc mb my_object_storage/nebula-br-bucket &&
        sleep 36000

I try to call with

backup full --meta "127.0.0.1:9559" --s3.endpoint "http://nginx:9000" --storage s3://nebula-br-bucket/ --s3.access_key=minioadmin --s3.secret_key=minioadmin --s3.region=default

also tried:

backup full --meta "127.0.0.1:9559" --s3.endpoint "http://nginx:9000" --storage s3://my_object_storage/nebula-br-bucket/ --s3.access_key=minioadmin --s3.secret_key=minioadmin --s3.region=default

but no matter what I give to storage it always says:

Put "http://nginx:9000/nebula-br-bucket/BACKUP_2023_11_25_21_03_22/BACKUP_2023_11_25_21_03_22.meta": dial tcp: lookup nginx on 127.0.0.11:53: no such host
...
{"level":"fatal","msg":"upload local tmp file to remote storage s3://nebula-br-bucket/BACKUP_2023_11_25_20_50_30/BACKUP_2023_11_25_20_50_30.meta failed: upload from /tmp/nebula-br/BACKUP_2023_11_25_20_50_30.meta to BACKUP_2023_11_25_20_50_30/BACKUP_2023_11_25_20_50_30.meta failed: RequestError: send request failed\ncaused by: Put \"http://nginx:9000/nebula-br-bucket/BACKUP_2023_11_25_20_50_30/BACKUP_2023_11_25_20_50_30.meta\": dial tcp: lookup nginx on 127.0.0.11:53: no such host","time":"2023-11-25T20:50:32.131Z"}

Could you help me how to specify storage?