[GH-ISSUE #797] [BUG] Misleading error message when git is not installed on periphery machine #3575

New Issue

GiteaMirror · 2026-04-13T15:05:25-05:00

GiteaMirror commented

2026-04-13 15:05:25 -05:00

Originally created by @simtel12 on GitHub (Sep 1, 2025).
Original GitHub issue: https://github.com/moghtech/komodo/issues/797

When the periphery machine does not have git installed, the error message is:

Failed to validate run directory on host after stack write (canonicalize error): 
No such file or directory (os error 2)

In this case, the periphery was running under systemd on multiple machines successfully, and while trying to onboard a new machine I got the above whenever I would try to run Pull Images for a stack. No other information in the logs, even at TRACE level.

Discovered the cause after reading the source and realizing that the error message is part of a path-validation after the git clone, and examining the stacks directory on the offending machine did not have a clone. So I'm guessing the command not found error is getting swallowed somewhere.

Environment:
Raspberry Pi 5, running dietpi
Periphery deployed via ansible
Systemd user service
Low-priviledge account komodo

Btw, I call it a misleading message because the only discussion about this error is in relation to docker and relative paths - which took me down the entirely wrong rabbit-hole. It is technically a correct error message (the run directory did not exist because the repo was not cloned).

Originally created by @simtel12 on GitHub (Sep 1, 2025). Original GitHub issue: https://github.com/moghtech/komodo/issues/797 When the periphery machine does not have `git` installed, the error message is: ``` Failed to validate run directory on host after stack write (canonicalize error): No such file or directory (os error 2) ``` In this case, the periphery was running under systemd on multiple machines successfully, and while trying to onboard a new machine I got the above whenever I would try to run `Pull Images` for a stack. No other information in the logs, even at TRACE level. Discovered the cause after reading the source and realizing that the error message is part of a path-validation _after_ the git clone, and examining the `stacks` directory on the offending machine did not have a clone. So I'm guessing the `command not found` error is getting swallowed somewhere. Environment: Raspberry Pi 5, running `dietpi` Periphery deployed via ansible Systemd user service Low-priviledge account `komodo` Btw, I call it a misleading message because the only discussion about this error is in relation to docker and relative paths - which took me down the entirely wrong rabbit-hole. It is *technically* a correct error message (the run directory did not exist because the repo was not cloned).

GiteaMirror added the bug label 2026-04-13 15:05:25 -05:00

GiteaMirror commented

2026-04-13 15:05:26 -05:00

@yujqiao commented on GitHub (Dec 2, 2025):

Ran into the same issue. I opened a PR to at least log the path so it's easier to pin down that the root cause is not cloned

@yujqiao commented on GitHub (Dec 2, 2025): Ran into the same issue. I opened a PR to at least log the path so it's easier to pin down that the root cause is not cloned

Sign in to join this conversation.

Branches Tags

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: github-starred/komodo#3575