TimeMachine Share not available

[andrema2]

Hi all

From time to time, almost daily my Macs are unable to perform backups to the TimeMachine share. It says disks are not available on Mac side.

If I try to manually perform a backup several times it sometimes works and other it doesn't. When it doesn't work I end up restarting the server and waiting for sometime which normally solves the issue.

The share seems to be correct in terms of the settings, I followed an instruction I found online.

I have the latest MacOS and latest FreeNAS-9.10.1-U2 (f045a8b)

I have a feeling it is a DNS or bonjour issue. Does anyone know how to solve it ?

Thanks

 

[Jann]

Hi,

I have a similar problem. My setup is still in testing, so changing is easy. Running latest 9.10. OS X not upgraded to Sierra yet. Both machines are in the same subnet.

The console/messages output on my Mac says it can't find the mDNS entry for the TM disk.

So far I've tried 2 things:
- reboot FreeNas. This solved the problem for some time (one or two backups)
- disabled local firewall on Mac (little snitch)

Still testing if it is stable now.
During troubleshooting I noticed several mDNSresponder messages that might indicate a problem.

I'm not very familiar with FreeBSD and how FreeNas manages services: How do I restart the mDNSresponder on FreeNas?

Jan





Sent from my iPad using Tapatalk

 

[Jann]

Hi,

I have a similar problem. My setup is still in testing, so changing is easy. Running latest 9.10. OS X not upgraded to Sierra yet. Both machines are in the same subnet.

The console/messages output on my Mac says it can't find the mDNS entry for the TM disk.

So far I've tried 2 things:
- reboot FreeNAS. This solved the problem for some time (one or two backups)
- disabled local firewall on Mac (little snitch)

Still testing if it is stable now.
During troubleshooting I noticed several mDNSresponder messages that might indicate a problem.

I'm not very familiar with FreeBSD and how FreeNAS manages services: How do I restart the mDNSresponder on FreeNAS?

Jan





Sent from my iPad using Tapatalk

I think I've found the problem:

Making changes in FreeNas via the web interface (e.g. adding/changing a user) triggers reload for several services including the mDNSresponder (mdnds). mdnds is restarted. Due to this restart the _afpovertcp.... mDNS entry gets lost / is not registered again. This entry is needed by the Macs to find the TimeMachine disk.

Changing a user caused the following entries in /var/log/daemon.log:

Code:

Oct 29 14:12:02 bigbird notifier: Performing sanity check on Samba configuration: OK
Oct 29 14:12:02 bigbird notifier: Reloading nmbd.
Oct 29 14:12:02 bigbird notifier: Reloading smbd.
Oct 29 14:12:02 bigbird notifier: Reloading winbindd.
Oct 29 14:12:02 bigbird notifier: Stopping mdnsd.
Oct 29 14:12:02 bigbird mDNSResponder: mDNSResponder (Engineering Build) (Oct  3 2016 06:58:50) stopping
Oct 29 14:12:02 bigbird mDNSResponder: mDNS_FinalExit failed to send goodbye for: 0000000801463580 01   25 _smb._tcp.local. PTR bigbird._smb._tcp.local.
Oct 29 14:12:02 bigbird mDNSResponder: mDNS_FinalExit failed to send goodbye for: 0000000801471580 01   32 _afpovertcp._tcp.local. PTR bigbird._afpovertcp._tcp.local.
Oct 29 14:12:02 bigbird mDNSResponder: mDNS_FinalExit failed to send goodbye for: 0000000801482580 01   27 _adisk._tcp.local. PTR bigbird._adisk._tcp.local.
Oct 29 14:12:02 bigbird mDNSResponder: mDNS_FinalExit failed to send goodbye for: 0000000801484580 01   33 _device-info._tcp.local. PTR bigbird._device-info._tcp.local.
Oct 29 14:12:02 bigbird mDNSResponder: mDNS_FinalExit failed to send goodbye for: 0000000801469580 01   25 _ssh._tcp.local. PTR bigbird._ssh._tcp.local.
Oct 29 14:12:02 bigbird mDNSResponder: mDNS_FinalExit failed to send goodbye for: 000000080147E580 01   30 _sftp-ssh._tcp.local. PTR bigbird._sftp-ssh._tcp.local.
Oct 29 14:12:02 bigbird mDNSResponder: mDNS_FinalExit failed to send goodbye for: 0000000801480580 01   27 _https._tcp.local. PTR bigbird._https._tcp.local.
Oct 29 14:12:02 bigbird netatalk[16034]: dnssd_clientstub DNSServiceProcessResult called with DNSServiceRef with no ProcessReply function
Oct 29 14:12:02 bigbird netatalk[16034]: dnssd_clientstub DNSServiceProcessResult called with DNSServiceRef with no ProcessReply function
Oct 29 14:12:02 bigbird netatalk[16034]: dnssd_clientstub DNSServiceProcessResult called with DNSServiceRef with no ProcessReply function
Oct 29 14:12:02 bigbird netatalk[16034]: dnssd_clientstub DNSServiceProcessResult called with DNSServiceRef with no ProcessReply function
Oct 29 14:12:02 bigbird netatalk[16034]: dnssd_clientstub DNSServiceProcessResult called with DNSServiceRef with no ProcessReply function
Oct 29 14:12:02 bigbird netatalk[16034]: dnssd_clientstub DNSServiceProcessResult called with DNSServiceRef with no ProcessReply function
Oct 29 14:12:02 bigbird netatalk[16034]: dnssd_clientstub DNSServiceProcessResult called with DNSServiceRef with no ProcessReply function
Oct 29 14:12:02 bigbird netatalk[16034]: dnssd_clientstub DNSServiceProcessResult called with DNSServiceRef with no ProcessReply function
Oct 29 14:12:02 bigbird netatalk[16034]: dnssd_clientstub DNSServiceProcessResult called with DNSServiceRef with no ProcessReply function
Oct 29 14:12:02 bigbird netatalk[16034]: dnssd_clientstub DNSServiceProcessResult called with DNSServiceRef with no ProcessReply function
Oct 29 14:12:02 bigbird notifier: Waiting for PIDS: 40321.
Oct 29 14:12:02 bigbird notifier: Starting mdnsd.
Oct 29 14:12:02 bigbird mDNSResponder: mDNSResponder (Engineering Build) (Oct  3 2016 06:58:50) starting
Oct 29 14:12:02 bigbird mDNSResponder:  13: Listening for incoming Unix Domain Socket client requests
Oct 29 14:12:02 bigbird mDNSResponder: mDNS_AddDNSServer: Lock not held! mDNS_busy (0) mDNS_reentrancy (0)
Oct 29 14:12:04 bigbird mDNSResponder: mDNS_Register_internal: ERROR!! Tried to register AuthRecord 0000000801406D60 bigbird.local. (Addr) that's already in the list
Oct 29 14:12:04 bigbird mDNSResponder: mDNS_Register_internal: ERROR!! Tried to register AuthRecord 0000000801407180 <ipv4-removed>.in-addr.arpa. (PTR) that's already in the list
Oct 29 14:12:04 bigbird mDNSResponder: mDNS_Register_internal: ERROR!! Tried to register AuthRecord 0000000801409D60 bigbird.local. (AAAA) that's already in the list
Oct 29 14:12:04 bigbird mDNSResponder: mDNS_Register_internal: ERROR!! Tried to register AuthRecord 000000080140A180 <link-local ipv6 removed>.ip6.arpa. (PTR) that's already in the list
Oct 29 14:12:04 bigbird mDNSResponder: mDNS_Register_internal: ERROR!! Tried to register AuthRecord 000000080140BD60 bigbird.local. (AAAA) that's already in the list
Oct 29 14:12:04 bigbird mDNSResponder: mDNS_Register_internal: ERROR!! Tried to register AuthRecord 000000080140C180 <global ipv6 removed>.ip6.arpa. (PTR) that's already in the list
Oct 29 14:18:37 bigbird notifier: Generating grub configuration file ...
Oct 29 14:18:38 bigbird notifier: done

After this the Mac can't find the time machine disk, messages output from disk:

Code:

29-10-16 14:28:39,307 com.apple.backupd[33864]: Timed out while resolving Bonjour destination: afp://tmgrover@bigbird._afpovertcp._tcp.local./tmgrover
29-10-16 14:28:39,317 com.apple.backupd[33864]: Starting automatic backup
29-10-16 14:28:39,326 com.apple.backupd[33864]: Attempting to soft mount network destination URL: afp://tmgrover@bigbird._afpovertcp._tcp.local./tmgrover
29-10-16 14:28:39,505 com.apple.backupd[33864]: Destination bigbird could not be found (url: afp://tmgrover@bigbird._afpovertcp._tcp.local./tmgrover destinationID: 0FEB4AB5-4392-48E0-97DC-D2A5EB8500B3)
29-10-16 14:28:39,523 com.apple.backupd[33864]: Backup failed with error 18: The backup disk could not be found.
29-10-16 14:28:39,548 com.apple.backupd[33864]: Starting automatic backup

This is solved by restarting the AFS service on FreeNAS. Apparently the AFS service registers the needed services with the mDNSResponder.
My solution for now is to stop and start the AFS service in the web interface:

Code:

Oct 29 14:55:35 bigbird cnid_metad[16037]: shutting down on SIGTERM
Oct 29 14:55:35 bigbird afpd[16036]: AFP Server shutting down
Oct 29 14:55:35 bigbird netatalk[16034]: Netatalk AFP server exiting
Oct 29 14:55:35 bigbird notifier: No matching processes were found
Oct 29 14:55:35 bigbird notifier: No matching processes were found
Oct 29 14:55:36 bigbird notifier: Starting netatalk.
Oct 29 14:55:37 bigbird netatalk[15858]: Netatalk AFP server starting
Oct 29 14:55:37 bigbird netatalk[15858]: Registered with Zeroconf
Oct 29 14:55:37 bigbird cnid_metad[15861]: CNID Server listening on localhost:4700
Oct 29 14:55:37 bigbird afpd[15860]: Netatalk AFP/TCP listening on 10.1.1.90:548
Oct 29 14:55:49 bigbird afpd[15896]: Login by tmgrover (AFP3.4)

I think 'Registered with Zeroconf' (mdns is part of zeroconf suite of protocols) registers the afptcpovertcp_... entry.
After this the Mac finds the Time Machine disk again:

Code:

29-10-16 14:55:48,601 com.apple.backupd[34032]: Starting manual backup
29-10-16 14:55:48,609 com.apple.backupd[34032]: Attempting to soft mount network destination URL: afp://tmgrover@bigbird._afpovertcp._tcp.local./tmgrover
29-10-16 14:55:49,649 com.apple.backupd[34032]: Mounted network destination at mount point: /Volumes/tmgrover using URL: afp://tmgrover@bigbird._afpovertcp._tcp.local./tmgrover

Am I correct? If yes, should this be considered a bug?

Jan

 

[Stux]

I would consider this a bug. If that is the case, you should file a bug report

 

[JJDuru]

I would consider this a bug. If that is the case, you should file a bug report

Yes, it is a bug and it's already filed and unsolved for almost two years:

freenas bug 7133:
Bug #7133: Time Machine fails trying to find disk until AFP is restarted ...

Moreover, appears that the status is "Cannot reproduce" while there are sufficient enough people hitting this bug over and over again.

Btw, I am using the latest stable freenas 9.10.1-U2 and I am hitting this bug myself as well.

 

[Jann]

Yes, it is a bug and it's already filed and unsolved for almost two years:

freenas bug 7133:
Bug #7133: Time Machine fails trying to find disk until AFP is restarted ...

Moreover, appears that the status is "Cannot reproduce" while there are sufficient enough people hitting this bug over and over again.

Btw, I am using the latest stable freenas 9.10.1-U2 and I am hitting this bug myself as well.

I've added some info to this bug report. Hopefully this helps to reproduce the problem.

For now I know what to do: restart AFS.

 

[jpaetzel]

I appreciate people making noise about this. I found the bug and fixed it. (not bad for a non-Mac user huh?)

 

[jpaetzel]

https://github.com/freenas/freenas/commit/0631525b22a95b6c36bdc6c94803312dd56e725b

It will be in FreeNAS 9.10.2

 

[Ericloewe]

https://github.com/freenas/freenas/commit/0631525b22a95b6c36bdc6c94803312dd56e725b

It will be in FreeNAS 9.2.10 9.10.2

Fixed that for you, that would've caused endless confusion among new users.

 

[NasFreeMePlease]

Time machines was working fine with 9.10.1-U4 (and has worked since I build freenas last year.) Earlier this week, I updated my freenas system to 9.10.2 and time machine stopped working. Seems the issue is the same with mDNS as discussed in this thread. Tried turning off and on AFP service and still not able to get it to work.

In addition, now SMB or AFP shares advertised through mDNS will show up in finder, but can not be connected to using the mDNS advertised share. The same mDNS errors noted in this thread are present when trying to connect to the advertised share. Connecting with the IP address to either protocol works fine.

 

[Dotty]

Do you have any Jail running?
Try turning all the Jails OFF if you can.
If that solve your problem, then go to your NIC and enable "Auto configure IPv6" (dont worry if you dont use IPv6)
Try then your Time Machine backup with your Jails ON and please report back.
;-)

 

[Todor Yankov]

Hi All,

I am new to FreeNAS and i have similar problem as described above.
A bit more for our setup:
I am responsible for setting up FreeNAS in my company and the network layout is as follow. We keep out wireless, LAN, Servers and Printers all on different VLANs, so the FreeNAS server is on different VLAN than the wired and wireless network.
I have created Time Machine share for Macbook backups (laptops are running latest OS X Sierra 10.12.4) and we are using the latest FreeNAS 9.10.2-U3 (e1497f269).
When the laptop is on the wired/wireless network i am able to map the share to finder using afp://IP address, but when i go to the Time Machine client the drive does not appear in the list of available locations to backup.
If i move the laptop on the same VLAN as the server everything work as it should be. I have tried restarting the AFP service on the server and all other workaround listed above, but non of them helps.
Bonjour is enable across those VLANs . No Jail running on the server.

The whole setup is still at a test stage so no problem of making any changes on the server.

Any help will be much appreciated.

Thanks

 

[Jann]

Hi All,

I am new to FreeNAS and i have similar problem as described above.
A bit more for our setup:
I am responsible for setting up FreeNAS in my company and the network layout is as follow. We keep out wireless, LAN, Servers and Printers all on different VLANs, so the FreeNAS server is on different VLAN than the wired and wireless network.
I have created Time Machine share for Macbook backups (laptops are running latest OS X Sierra 10.12.4) and we are using the latest FreeNAS 9.10.2-U3 (e1497f269).
When the laptop is on the wired/wireless network i am able to map the share to finder using afp://IP address, but when i go to the Time Machine client the drive does not appear in the list of available locations to backup.
If i move the laptop on the same VLAN as the server everything work as it should be. I have tried restarting the AFP service on the server and all other workaround listed above, but non of them helps.
Bonjour is enable across those VLANs . No Jail running on the server.

The whole setup is still at a test stage so no problem of making any changes on the server.

Any help will be much appreciated.

Thanks

Todor,

Time Machine depends on mDNS/bonjour to find the shares (not just the host names). mDNS is a multicast protocol. If your server and client are on different VLANs they are probably also on different IP subsets and the traffic is going via an IP router.

To make your situation work you need a way to propagate/proxy the mDNS traffic to your different subsets/VLANs. How this is done depends on your type of router.

Some open source routers have a avahi daemon, others have support for mDNS or bonjour in proprietary implementations. Enable this and your problem might be solved.

If your router doesn't support this, you can also install mDNS on any server connected to both VLANs.

Jan

 

[Todor Yankov]

Jann,

Thanks for the reply.
Actually i sorted the issue by myself. There was a problem with my network. I had to enable AFP service to go trough the VLANs.

Thanks

 

[pixel8383]

After this the Mac can't find the time machine disk, messages output from disk:

Same issue to my configuration. An AFP restart solve the problem but it's just a workaround, not a solution. I am running FreeNAS 11.1-U5

 

[Dotty]

Same issue to my configuration. An AFP restart solve the problem but it's just a workaround, not a solution. I am running FreeNAS 11.1-U5

See my post above

 

[pixel8383]

Do you have any Jail running?
Try turning all the Jails OFF if you can.
If that solve your problem, then go to your NIC and enable "Auto configure IPv6" (don't worry if you don't use IPv6)
Try then your Time Machine backup with your Jails ON and please report back.
;-)

Hello Dotty, I will have a try. Which NIC do you refer to? The jail's one? The macOS's one? The FreeNAS's one?

 

[b88]

I've had similar kind of issues for a long time and couldn't figure out how to make afp shares and time machine work reliably. More specifically:
freenas server does not show up in finder, time machine only works sometimes, connecting direct to server via cmd+k always works.

However i stumbled upon this:
https://forums.freenas.org/index.php?threads/interference-between-jail-and-afp-share.28065/

It suggests that in order to make afp work properly, one must enable ipv6 autoconfigure (Network->Interfaces) for the interface on which the afp is advertising.

For me! Blistering fast connects and always a server symbol in the finder sidebar. However I am no longer able to run my plex jail (not plugin). When i try to start the jail I get a message at the top of the GUI saying something went wrong, but the jail starts anyway and the afp shares no longer work. It also seems that i need to bind the afp service to both the ipv4 and ipv6 addresses of the active interface (only use one interface).

This might solve your issue with time machine. However you still have to choose between using afp service or using jails.

PS: It seems that something in finder breaks when the macbook has been in sleep mode and sometimes it can not connect. Then a restart of finder does the trip. Will do some more testing when it comes to sleep and afp.

 

[pixel8383]

Hello b88, thanks for your reply. I have readen the post you linked however I did not undestood how to enable ipv6 to my primary (and only) NIC. The default one, as far I know, is not listed into Network > Interfaces so I cannot enable it using the GUI as I guess. Where should I look for?
Thanks.

 

[b88]

Hello b88, thanks for your reply. I have readen the post you linked however I did not undestood how to enable ipv6 to my primary (and only) NIC. The default one, as far I know, is not listed into Network > Interfaces so I cannot enable it using the GUI as I guess. Where should I look for?
Thanks.

Click on the "Add Interface" button, to add your one and only NIC. If non is listed, freenas chooses one by itself. After adding an interface you can enable ipv6.