recon middleware complains too loudly about missing recon files

[clayg]

It's probably not that big of deal for most recon endpoints because the account/container/object.recon files are created pretty quickly when services come online.

But the drive.recon file may never exist if you're not using swift-drive-audit [1] - so you get a traceback everytime you hit the /driveaudit endpoint

ubuntu@saio:~$ curl -s http://localhost:6010/recon/driveaudit | python -m json.tool
{
    "drive_audit_errors": null
}
ubuntu@saio:~$ tail /var/log/syslog -n 2
Dec 1 09:31:52 saio object-6010: Error reading recon cache file: #012Traceback (most recent call last):#012 File "/vagrant/swift/swift/common/middleware/recon.py", line 81, in _from_recon_cache#012 with openr(cache_file, 'r') as f:#012IOError: [Errno 2] No such file or directory: '/var/cache/swift/node1/drive.recon'
Dec 1 09:31:52 saio object-6010: STDERR: 127.0.0.1 - - [01/Dec/2016 09:31:52] "GET /recon/driveaudit HTTP/1.1" 200 136 0.002146

It was reported on the ML [2] the noisiness of the logged trackebacks [3] is problematic

We should reduce/remove those log messages.

1. http://docs.openstack.org/developer/swift/admin_guide.html?highlight=drive#detecting-failed-drives
2. http://lists.openstack.org/pipermail/openstack/2016-December/018065.html
3. https://github.com/openstack/swift/blob/43a175ebd284e825c0f6e5b79a23d8a32f62326e/swift/common/middleware/recon.py#L85