swift-recon-cron lock dir not removed on failure

[Christopher H. Laco]

Under certain circumstances, if get_async_count or dump_cron_cache calls fail, the lock_dir is not cleaned up.

You can simulate this by running this cron on a machine without /srv/node, which will yield the following exception:

Traceback (most recent call last):
  File "/usr/bin/swift-recon-cron", line 78, in <module>
    main()
  File "/usr/bin/swift-recon-cron", line 70, in main
    dump_recon_cache({'async_pending': asyncs}, cache_file, logger)
UnboundLocalError: local variable 'asyncs' referenced before assignment

All subsequent executions of swift-recon-cron will fail with the following error until the lock_dir is removed:

[Errno 17] File exists: '/var/lock/swift/swift-recon-object-cron'
2013 Nov 15 16:44:02 storage-server-1 [Errno 17] File exists: '/var/lock/swift/swift-recon-object-cron'

Error handling should probably be a bit more robust to ensure the lock_dir always gets removed. Patch forthcoming.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/56670