[v2] telepresence volume mount is empty

[vicnoah]

Local:

Pod:

You can see that there is a file in the pod for the mounted volume, but it is empty after the local mount.

[ipleten]

Might be related to #1930.
Can you check if there is something related to sshfs in connector.log

[vicnoah]

Might be related to #1930.
Can you check if there is something related to sshfs in connector.log

I don't know how to view the logs of sshfs.

[ipleten]

Might be related to #1930.
Can you check if there is something related to sshfs in connector.log

I don't know how to view the logs of sshfs.

This is a telepresence's log. Its location depends on OS you are using.
assuming this is *nix it might be somewhere it $HOME/telepresence/
you can use

find /${HOME} -name 'connector.log'
to find it

[vicnoah]

Might be related to #1930.
Can you check if there is something related to sshfs in connector.log

I don't know how to view the logs of sshfs.

This is a telepresence's log. Its location depends on OS you are using.
assuming this is *nix it might be somewhere it $HOME/telepresence/
you can use

find /${HOME} -name 'connector.log'
to find it

This is the content of the log.
connector.log

[mentos1386]

I had similar issue, solution was to do echo user_allow_other >> /etc/fuse.conf

[vicnoah]

I had similar issue, solution was to do echo user_allow_other >> /etc/fuse.conf

Thanks, that's right.

[TavaresDylan]

I had similar issue, solution was to do echo user_allow_other >> /etc/fuse.conf

Thanks , same problem same solution too

[rotty3000]

FYI another +1 on the issue and the proposed solution.

However, another related issue really troubled me. It was not clear to me that if you specify a mount, it must NOT exist.

In other words, if you specify a mount point of say /tmp/ and that exists, it would not mount for me at all, even with the fuse permission. I had to specify a mount path that didn't previously exist, then telepresence would create the dir, and mount the files in it (and also delete the new dir during shutdown).

Secondly, and this may be more specific to my app, but the path needed to be absolute. A relative path was causing issues for me (again this may have been my apps poor handling of the path.)

Thx for the really nice tool.

[SomaticIT]

Hello,

I got a similar issue on WSL 2.
/etc/fuse.conf already contains user_allow_other.

Here is the content of connector.log:

2021-10-12 01:02:32.5157 info    Logging at this level "info"
2021-10-12 01:02:32.5166 info    ---
2021-10-12 01:02:32.5167 info    Telepresence Connector v2.4.4 (api v3) starting...
2021-10-12 01:02:32.5167 info    PID is 11072
2021-10-12 01:02:32.5167 info
2021-10-12 01:02:32.5168 info    connector/server-grpc : gRPC server started
2021-10-12 01:02:32.7325 info    connector/background-init : Connecting to daemon...
2021-10-12 01:02:32.7335 info    connector/background-init : Connecting to k8s cluster...
2021-10-12 01:02:33.0081 info    connector/background-init : Server version v1.20.5
2021-10-12 01:02:33.0082 info    connector/background-init : Context: **********
2021-10-12 01:02:33.0082 info    connector/background-init : Server: https://**********
2021-10-12 01:02:33.0082 info    connector/background-init : Connected to context ********** (https://**********)
2021-10-12 01:02:33.3306 info    connector/background-init : Connecting to traffic manager...
2021/10/12 01:02:33 Patching synced Namespace **********
2021-10-12 01:02:33.3307 info    connector/background-init : Waiting for TrafficManager to connect
2021-10-12 01:02:33.4512 info    connector/background-manager : Existing Traffic Manager not owned by cli or does not need upgrade, will not modify
2021-10-12 01:02:35.0943 info    connector/server-grpc/conn=1/CreateIntercept-3 : Waiting for agent for Deployment **********
2021-10-12 01:02:35.0947 info    connector/server-grpc/conn=1/CreateIntercept-3 : Agent found or created for Deployment **********
2021-10-12 01:02:35.1517 info    connector/background-manager/intercept-port-forward/100.65.132.179:41657 : Mounting file system for intercept "**********" at "/tmp/telfs-2451038024"
2021/10/12 01:07:33 Patching synced Namespace **********
2021-10-12 01:09:34.2095 info    connector/background-manager/intercept-port-forward : Terminating forwards for 100.65.132.179
2021-10-12 01:09:34.2097 info    connector/background-manager/intercept-port-forward : Removed file system mount "/tmp/telfs-2451038024"
2021-10-12 01:09:34.2750 info    connector/background-k8swatch:shutdown_logger : shutting down (gracefully)...
2021-10-12 01:09:34.2751 info    connector/background-manager:shutdown_logger : shutting down (gracefully)...
2021-10-12 01:09:34.2752 info    connector:shutdown_logger : shutting down (gracefully)...
2021-10-12 01:09:34.2752 info    connector/background-systema:shutdown_logger : shutting down (gracefully)...
2021-10-12 01:09:34.4759 info    connector/server-grpc/conn=2:shutdown_logger : shutting down (gracefully)...

[SomaticIT]

Any news?

[rotty3000]

just in case, @SomaticIT and you tried with a directory which does not exist before mounting? (see my comment)

[SomaticIT]

Yes, i retried to be sure.

[SomaticIT]

For you information, the issue was coming from a misconfiguration of the podCIDRStrategy.
We fixed the issue by installing the traffic-manager manually with the podCIDRStrategy: coverPodIPs value.
The issue is covered here: #2135.