Multinode ECS Install fails

Hello Holger,

I faced the same issue and decided to try correcting the script code to solve it.

I did solve this issue with the 'DockerCommandLineFlags' variable not defined, and some other minor issues, but at the end there is still something missing in this script as I'm stuck in the authentication phase, where the script tries to connect to the ECS service inside the container but id cannot even after 30 minutes trying...

[22/Apr/2016 12:53:03] INFO [root:596] Problem reaching authentication server. Retrying shortly.

Executing getAuthToken: curl -i -k https://10.5.124.78:4443/login -u root:ChangeMe

  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current

                                 Dload  Upload   Total   Spent    Left  Speed

  0     0    0     0    0     0      0      0 --:--:-- --:--:-- --:--:--     0curl: (7) Failed connect to 10.5.124.78:4443; Connection refused

[22/Apr/2016 12:53:33] INFO [root:596] Problem reaching authentication server. Retrying shortly.

[22/Apr/2016 12:53:33] CRITICAL [root:599] Authentication service not yet started.

[22/Apr/2016 12:53:33] INFO [root:704] Step 1 Completed.  Navigate to the administrator website that is available from any of the ECS data nodes.         The ECS administrative portal can be accessed from port 443. For example: https://ecs-node-external-ip-address.         The website may take a few minutes to become available.

In case you want to try, I attached my script with the corrections mentioned above.

ps: I'm not a developer or anything like it, just a user that wanted to experiment with ECS and faced the same issue..

Regards,

Caio

Hello Aaron,

I tried the latest version on github today but I found a strange error that mentioned the ecsstandalone container name, instead of the ecsmultinode container name. I found the reference on line 553 of the step1 script:

os.system("docker "+' '.join(DockerCommandLineFlags)+" exec -t -i ecsstandalone bash -c \"cd /tmp/kennethreitz-requests-* && python setup.py install\"")

Even after correcting this command I still can't get access to ECS service even after 30 minutes of waiting.... Do you know if there is some additional work to be done in order to get this working?

btw: I have shutdown the firewalld service and it still does not work... the most strange thing is that if I decide to run the script for the single node ECS deployment on the same host, it works like a charm....

Thanks

Hi Aaron

step1 script worked. step2 failed with:

[ecs@ecsnode1 ecs-multi-node]$ sudo python step2_object_provisioning.py --ECSNodes=192.168.1.22 192.168.1.23 192.168.1.24 192.168.1.25 --Namespace=mnns1 --ObjectVArray=mnova1 --ObjectVPool=mnovp1 --UserName=emccode --DataStoreName=mnds1 --VDCName=mnvdc1 --MethodName=

[sudo] password for ecs:

Executing getAuthToken: curl -i -k https://192.168.1.22:4443/login -u root:ChangeMe

  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current

                                 Dload  Upload   Total   Spent    Left  Speed

100    93  100    93    0     0    169      0 --:--:-- --:--:-- --:--:--   170

Auth Token  BAAcZHFkRnRtaGlaNG1ocDlycHlxT2ZMdEZLbHo0PQMAjAQASHVybjpzdG9yYWdlb3M6VmlydHVhbERhdGFDZW50ZXJEYXRhOmJkM2NlMzNlLWM2NjYtNDM4Yi04ZjY3LTIzYzA4MGMxMmYzMQIADTE0NjE1ODg3NTQ0ODgDAC51cm46VG9rZW46OTQwYWQ3YzQtNWZhZS00ZTZlLWFjYmMtMDM0ODM4OWU0MDNlAgAC0A8=

ECSNodes: 192.168.1.22

Traceback (most recent call last):

  File "step2_object_provisioning.py", line 360, in

    main(sys.argv[1:])

  File "step2_object_provisioning.py", line 279, in main

    print("ObjectVArray: %s" %ObjectVArray)

UnboundLocalError: local variable 'ObjectVArray' referenced before assignment

[ecs@ecsnode1 ecs-multi-node]$

Thanks a lot, Holger

Holger,

I believe it's a result of your formatting; you should use a comma-delimited list rather than a space-delimited one for ECSNodes, e.g.

[ecs@ecsnode1 ecs-multi-node]$ sudo python step2_object_provisioning.py --ECSNodes=192.168.1.22,192.168.1.23,192.168.1.24,192.168.1.25 --Namespace=mnns1 --ObjectVArray=mnova1 --ObjectVPool=mnovp1 --UserName=emccode --DataStoreName=mnds1 --VDCName=mnvdc1 --MethodName=

Honestly not sure what Calo is seeing if his environment is working fine with single-node but failing authentication on multi-node for the same host; there shouldn't be additional setup necessary by comparison. Other developers with possible insight would be welcomed here.

Hi Aaron

Thanks for the reply and correction. Sorry for the trouble.

I will retry, even on my ssd esx it will take a short while. Then I'll update you.

As for the failing authentication, this took a long while on my machine to complete. But it finished.

I will keep you posted, Holger