search

bsdci / libioc

A Python library to manage jails with ioc{age,ell}
https://bsd.ci/libioc
Other
38 stars 11 forks source link

`ioc destroy -f` should call `ioc stop -f` #732

Open igalic opened 5 years ago

igalic commented 5 years ago 
root@webserver /u/l/s/i/.libioc# ioc destroy -f blog
[-] ZFSDatasetDestroy@zroot/ioc/jails/blog: ...
Traceback (most recent call last):
  File "/usr/local/bin/ioc", line 10, in <module>
    sys.dd:exit(cli())
  File "/usr/local/lib/python3.6/site-packages/click/core.py", line 764, in __call__
    return self.main(*args, **kwargs)
  File "/usr/local/lib/python3.6/site-packages/click/core.py", line 717, in main
    rv = self.invoke(ctx)
  File "/usr/local/lib/python3.6/site-packages/click/core.py", line 1137, in invoke
    return _process_result(sub_ctx.command.invoke(sub_ctx))
  File "/usr/local/lib/python3.6/site-packages/click/core.py", line 956, in invoke
    return ctx.invoke(self.callback, **ctx.params)
  File "/usr/local/lib/python3.6/site-packages/click/core.py", line 555, in invoke
    return callback(*args, **kwargs)
  File "/usr/local/lib/python3.6/site-packages/click/decorators.py", line 17, in new_func
    return f(get_current_context(), *args, **kwargs)
  File "/usr/local/lib/python3.6/site-packages/ioc_cli/destroy.py", line 125, in cli
    ctx.parent.print_events(item.destroy())
  File "/usr/local/lib/python3.6/site-packages/ioc_cli/__init__.py", line 91, in print_events
    for event in generator:
  File "/usr/local/src/ioc/.libioc/libioc/Jail.py", line 1177, in destroy
    raise e
  File "/usr/local/src/ioc/.libioc/libioc/Jail.py", line 1174, in destroy
    self.zfs.delete_dataset_recursive(self.dataset)
  File "/usr/local/src/ioc/.libioc/libioc/ZFS.py", line 96, in delete_dataset_recursive
    self.delete_dataset_recursive(child)
  File "/usr/local/src/ioc/.libioc/libioc/ZFS.py", line 101, in delete_dataset_recursive
    dataset.umount()
  File "libzfs.pyx", line 2547, in libzfs.ZFSDataset.umount
libzfs.ZFSException: Device busy
root@webserver /u/l/s/i/.libioc#

now let's call stop -f, the destroy -f again:

root@webserver /u/l/s/i/.libioc# ioc stop -f blog
[+] JailStop@blog: OK [0.096s]
  [+] JailHookPrestop@blog: SKIPPED [0.001s]
  [+] JailHookStop@blog: not running [0.0s]
  [+] JailRemove@blog: SKIPPED [0.001s]
  [+] JailHookPoststop@blog: SKIPPED [0.001s]
  [+] MountFstab@blog: OK [0.001s]
  [+] BasejailStorageConfig@blog: OK [0.019s]
  [+] TeardownSystemMounts@blog: OK [0.002s]
  [+] JailResourceLimitAction@blog: SKIPPED [0.001s]
blog stopped
root@webserver /u/l/s/i/.libioc# ioc destroy -f blog
[+] ZFSDatasetDestroy@zroot/ioc/jails/blog: OK [0.1s]
/iocage/jails/blog destroyed
root@webserver /u/l/s/i/.libioc#

works. imo, destroy -f should work the same.

urosgruber commented 5 years ago

Agree, this would ease thing a bit.

gronke commented 5 years ago

@igalic to reproduce your case, the jail must have failed being stopped manually (without force). This means that the JID has been already removed, but the poststop task did not succeed.

I wonder if it would make sense to always force stop and remove the option to stop without that behavior. libioc without a daemon is stateless and we should never have jails in zombie state.

If that becomes the only way to stop, the problem raised in this Issue, would also disappear. What do you say?

igalic commented 4 years ago

either that, or we could make stop without -f block until it's done

of course that could hang "forever" so it would be nice if it was cancellable