where $SA is a storage account (either one of the five created by the Docker4Azure ARM template or any other Storage Account that you create/manage independently) and $SA_KEY is the access key for the storage account.
Running the command @ddebroy provided to manually install cloudstore worked properly. However, it is not practical as it needs to be installed on all workers and will need to be installed on new machines when scaling up the VM scaling set.
on one of the nodes where Cloudstor is in the disabled state and see if there are any errors?
I just tried the West Europe region and found Cloudstor installed there fine.
@agilyslaurent you can issue the command with a utility we have called swarm-exec in Docker4Azure nodes and that will make sure the command is executed in all nodes of the swarm as well as any new nodes that come up. However I would like to get some info on any errors that pop up in the /var/log/docker.log regarding Cloudstor installation issues.
@ddebroy
I just executed the command grep cloudstor /var/log/docker.log and it returns nothing.
To clarify, cloudstor is not in the disabled state : it is not installed at all (docker plugin ls returns no rows).
Also, we are creating the swarm via docker cloud swarm mode, not in the azure market place. Like I said earlier, maybe it is a problem with Docker Cloud rather than the Azure for Docker image?
Thank you for pointing out the swarm-exec. We will try it on our swarm.
@agilyslaurent Thanks for clarifying the bit about the fact that you are initializing through Docker Cloud. Cloudstor wonât get installed for swarms created by Docker Cloud by default at the moment. You will need to install cloudstor separately through swarm-exec and passing it the docker plugin install .. command line. We will address this shortly.
I had confused your message with the report about Cloudstor showing up disabled from @assurehedgedoc.
plugin shows disabled. I tied to install this manually via the command you suggested, I still get
Error response from daemon: dial unix /run/docker/plugins/b3e20d266064e51bff1585454ab254fdbcad57b39631195e3ec4732e2d38a9de/cloudstor.sock: connect: no such file or directory
logs:
Jul 17 03:24:32 moby root: time=â2017-07-17T03:24:32.609678200Zâ level=debug msg=âTrying to pull docker4x/cloudstor from https://registry-1.docker.io v2â
Jul 17 03:24:35 moby root: time=â2017-07-17T03:24:35.224390700Zâ level=debug msg=âPulling ref from V2 registry: docker4x/cloudstor:17.06.0-ce-azure1â
Jul 17 03:24:37 moby root: time=â2017-07-17T03:24:37.570243900Zâ level=debug msg=âCalling POST /v1.30/plugins/cloudstor:azure/setâ
Jul 17 03:24:37 moby root: time=â2017-07-17T03:24:37.600925700Zâ level=debug msg=âCalling POST /v1.30/plugins/cloudstor:azure/enable?timeout=0â
Jul 17 03:24:50 moby root: time=â2017-07-17T03:24:50.224349400Zâ level=debug msg=âerror net dialing plugin: dial unix /run/docker/plugins/b3e20d266064e51bff1585454ab254fdbcad57b39631195e3ec4732e2d38a9de/cloudstor.sock: connect: no such file or directoryâ
Jul 17 03:24:50 moby root: time=â2017-07-17T03:24:50.551648400Zâ level=debug msg=âerror net dialing plugin: dial unix /run/docker/plugins/b3e20d266064e51bff1585454ab254fdbcad57b39631195e3ec4732e2d38a9de/cloudstor.sock: connect: no such file or directoryâ
I am having the same issue as @andrewroth. I am not seeing the Cloudstor plugin on a newly deployed Docker Swarm (deployed via Docker Cloud using Microsoft Azure as service provider).
@ddebroy, could you provide some information on this issue?
Error response from daemon: dial unix /run/docker/plugins/3e5a75a71dc44d6dc93a8e4551434d5938147df069f763fd6761e12c69974ef4/cloudstor.sock: connect: no such file or directory
Is it related to the hosts configuration in daemon.json? Or maybe socket has to be exposed to the cloudstor container somehow?
It works, but! when i have 2 service containers running on the same node that uses the same mount, i am getting an error
starting container failed: error while mounting volume '/var/lib/docker/plugins/1165f4763631bcbebedf59d107f99b9013dc2c210db388172c8815fa61eacba3/rootfs': VolumeDriver.Mount: mount failed: exit status 32
output="mount error(16): Resource busy\nRefer to the mount.cifs(8) manual page (e.g. man mount.cifs)\nmount: mounting //hashspotprodmachinesdata.file.core.windows.net/6518c5505abe4a487674c76b823804da on /mnt/cloudstor/applogs failed: Resource busy\n""
that the cloudstor supports such situation when 2 containers running on the same node with same mapping?
(again - if i will have 2 nodes each running one container using cloudstor it will work,
but if i will have 2 nodes and one of them tries to run to containers with same cloudstor mapping - it will fail).
Iâm still having the same issue with version âdocker4x/cloudstor:18.06.0-ce-azure1â of the plugin:
Error response from daemon: dial unix /run/docker/plugins/b523465c05712da9213a8713ac41067be2365d425b77f68a442486fde2284ad9/cloudstor.sock: connect: no such file or directory
I can confirm what @harcher81 said, if the storage account has the âSecure transfer requiredâ setting enabled, I get the âError response from daemon: dial unix /run/docker/plugins//cloudstor.sock: connect: no such file or directoryâ message when trying to install or enable the plugin. After disabling this setting in the storage account, reinstalling the plugin immediately worked.