Has 9.1.2.6 broken Time Machine support?

[adrianwi]

I upgraded to 9.2.1.6 over the weekend and have been having lots of problems with Time Machine not running with an error that device is already in use.

It's not, as it's a dataset used specifically for backing up that Mac so it looks like its not completing correctly. If I switch off AFP and switch back on, it will work again for a day or two and then the error returns.

Anyone else experiencing this?

 

[fracai]

I'm also experiencing the "already in use" error, but I'm pretty sure I saw the same in an earlier version as well (9.1.2.3?).

I wrote a script that uses the API to turn AFP off and back on that I can post later tonight.

 

[fracai]

Code:

#!/usr/bin/env python

import requests, json, time

password='root_password'
host='freenas_host'

login = ('root', password)
headers = {'Content-Type': "application/json"}
url = 'http://'+host+'/api/v1.0/services/services/afp/'

result = requests.put(
    url,
    headers=headers,
    auth=login,
    data=json.dumps({'srv_enable': 'False'})
)

print result.json()
time.sleep(1)

result = requests.put(
    url,
    headers=headers,
    auth=login,
    data=json.dumps({'srv_enable': 'True'})
)

print result.json()

 

[DaPlumber]

Usually this happens for me when there's a disruption to a TM session network/power/whatever failure and TM doesn't release the "lock" it has on the TM volume. Rebooting netatalk (turn AFP off and on again as above) usually fixes the issue. If it doesn't them the TM volume is probably damaged and fixing that is a whole 'nother exercise. If this is occurring frequently you should probably go looking for what's disrupting the TM session, for me it's usually a network issue. I've had the exact same issue with Time Capsules too, except you have to reboot the whole TC which can be a PITA.