juju status results in unexpected fault address on arm64 using/ local provider
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
juju-core |
Invalid
|
High
|
Unassigned |
Bug Description
JuJu Version:
juju-core:
Installed: 1.17.7-0ubuntu1
Candidate: 1.17.7-0ubuntu1
After running juju bootstrap against the local provider, The bootstrap appeared to complete successfully. I did not see any lxc containers available after the bootstrap, nor did /var/log/lxc contain any logs. The /var/log/
Steps to reproduce on arm64:
1.) Install juju-local
2.) juju init
3.) juju switch local
4.) juju bootstrap --show-log --debug (wait for completion)
5.) juju status
sfeole@am1:~$ juju status
unexpected fault address 0x1
fatal error: fault
goroutine 7 [running]:
panic during panic
followed by,
sfeole@am1:~$ juju status
panic: runtime error: invalid memory address or nil pointer dereference
[signal 0xb code=0x1 addr=0x0]
goroutine 7 [running]:
goroutine 7 [running]:
goroutine 7 [running]:
goroutine 7 [running]:
created by launchpad.
goroutine 1 [chan receive]:
launchpad.
launchpad.
juju.newAPIFrom
juju.newAPIClient
launchpad.
main.Run.
main.Run.
launchpad.
launchpad.
main.Main
main.main
goroutine 3 [syscall]:
goroutine in C code; stack unavailable
goroutine 6 [select]:
launchpad.
parallel.$nested3
created by launchpad.
goroutine 11 [finalizer wait]:
tags: | added: local-provider status |
Changed in juju-core: | |
status: | New → Triaged |
importance: | Undecided → High |
Changed in juju-core: | |
milestone: | none → 1.19.0 |
This isn't actually a bug in Juju. The gcc Go compiler has issues on ARM which are actively being worked on.
eg
- autoconf is incorrectly turning on the -fsplit-stack option for gccgo
- the stack size is set too small
Marking as Invalid since the root cause lies outside Juju.