Updater quarantines bad asyncs to wrong policy data dir

[clayg]

When working on an out-of-tree tool to inspect some unprocessible async updates we borrowed some updater quarantine code and I noticed in review that it wasn't using the policy to create the quarantine directory which seemed un-idiomatic. I think this behavior of the updater on master is wrong:

vagrant@saio:~$ echo 'test1' > test
vagrant@saio:~$ swift upload test test -H 'x-storage-policy: ec'
test
vagrant@saio:~$ swift-init stop container-server -c 1
Signal container-server pid: 36926 signal: Signals.SIGTERM
container-server (36926) appears to have stopped
vagrant@saio:~$ swift-init stop container-server -c 2
Signal container-server pid: 36927 signal: Signals.SIGTERM
container-server (36927) appears to have stopped
vagrant@saio:~$ echo 'test2' > test
vagrant@saio:~$ swift upload test test
test
vagrant@saio:~$ find /srv/node*/sdb*/async*
/srv/node2/sdb6/async_pending-1
/srv/node2/sdb6/async_pending-1/35a
/srv/node2/sdb6/async_pending-1/35a/a161102fba1710ef912af194b8d4635a-1692886411.39630
/srv/node3/sdb3/async_pending-1
/srv/node3/sdb3/async_pending-1/35a
/srv/node3/sdb3/async_pending-1/35a/a161102fba1710ef912af194b8d4635a-1692886411.39630
vagrant@saio:~$ : > /srv/node2/sdb6/async_pending-1/35a/a161102fba1710ef912af194b8d4635a-1692886411.39630
vagrant@saio:~$ swift-init container-server restart
Signal container-server pid: 36928 signal: Signals.SIGTERM
Signal container-server pid: 36929 signal: Signals.SIGTERM
container-server (36928) appears to have stopped
container-server (36929) appears to have stopped
WARNING: Unable to modify max process limit. Running as non-root?
Starting container-server...(/etc/swift/container-server/1.conf.d)
Starting container-server...(/etc/swift/container-server/2.conf.d)
Starting container-server...(/etc/swift/container-server/3.conf.d)
Starting container-server...(/etc/swift/container-server/4.conf.d)
vagrant@saio:~$ swift-init object-updater once -nv
WARNING: Unable to modify max process limit. Running as non-root?
Running object-updater once...(/etc/swift/object-server/1.conf.d)
Running object-updater once...(/etc/swift/object-server/2.conf.d)
Running object-updater once...(/etc/swift/object-server/3.conf.d)
Running object-updater once...(/etc/swift/object-server/4.conf.d)
...
object-6020: ERROR Pickle problem, quarantining /srv/node2/sdb6/async_pending-1/35a/a161102fba1710ef912af194b8d4635a-1692886411.39630: #012Traceback (most recent call last):#012 File "/vagrant/swift/swift/obj/updater.py", line 430, in _load_update#012 return pickle.load(open(update_path, 'rb'))#012EOFError: Ran out of input
...
vagrant@saio:~$ find /srv/node*/sdb*/quar*
/srv/node2/sdb6/quarantined
/srv/node2/sdb6/quarantined/objects
/srv/node2/sdb6/quarantined/objects/a161102fba1710ef912af194b8d4635a-1692886411.39630

I think I'd actually expect these files to get moved into a path under /quarantined that matches their original path on disk, i.e.

/srv/node2/sdb6/quarantined/async_pending-1/35a/a161102fba1710ef912af194b8d4635a-1692886411.39630