macOS: Sierra

 

I can see the AFP share, but can not connect.

I can see and browse all SMB shares just fine.

 

Following log:

Jan 6 09:31:01 nerdyRAID avahi-daemon[19967]: Files changed, reloading.
Jan 6 09:31:01 nerdyRAID avahi-daemon[19967]: Service group file /services/afp.service changed, reloading.
Jan 6 09:31:02 nerdyRAID avahi-daemon[19967]: Service "nerdyRAID-AFP" (/services/afp.service) successfully established.
Jan 6 09:32:42 nerdyRAID emhttp: shcmd (43338): mkdir '/mnt/user/macUP' |& logger
Jan 6 09:32:42 nerdyRAID emhttp: shcmd (43339): chmod 0777 '/mnt/user/macUP'
Jan 6 09:32:42 nerdyRAID emhttp: shcmd (43340): chown 'nobody':'users' '/mnt/user/macUP'
Jan 6 09:34:02 nerdyRAID avahi-daemon[19967]: Files changed, reloading.
Jan 6 09:34:02 nerdyRAID avahi-daemon[19967]: Service group file /services/afp.service changed, reloading.
Jan 6 09:34:03 nerdyRAID avahi-daemon[19967]: Service "nerdyRAID-AFP" (/services/afp.service) successfully established.
Jan 6 09:34:46 nerdyRAID cnid_dbd[17084]: error opening DB environment after recovery: No such device
Jan 6 09:34:46 nerdyRAID cnid_dbd[17084]: Failed to open CNID database for volume "macUP"
Jan 6 09:34:46 nerdyRAID cnid_dbd[17084]: Recreated CNID BerkeleyDB databases of volume "macUP"
Jan 6 09:34:46 nerdyRAID cnid_dbd[17084]: error opening DB environment after recovery: No such device
Jan 6 09:34:46 nerdyRAID cnid_dbd[17084]: open_db() failed: No such device
Jan 6 09:34:46 nerdyRAID cnid_dbd[17084]: reinit_db() failed: No such device
Jan 6 09:34:46 nerdyRAID afpd[17065]: read: Connection reset by peer
Jan 6 09:34:46 nerdyRAID cnid_dbd[17109]: error opening DB environment after recovery: No such device
Jan 6 09:34:46 nerdyRAID cnid_dbd[17109]: Failed to open CNID database for volume "macUP"
Jan 6 09:34:46 nerdyRAID cnid_dbd[17109]: Recreated CNID BerkeleyDB databases of volume "macUP"
Jan 6 09:34:46 nerdyRAID cnid_dbd[17109]: error opening DB environment after recovery: No such device
Jan 6 09:34:46 nerdyRAID cnid_dbd[17109]: open_db() failed: No such device
Jan 6 09:34:46 nerdyRAID cnid_dbd[17109]: reinit_db() failed: No such device
Jan 6 09:34:46 nerdyRAID afpd[17065]: read: Connection reset by peer
Jan 6 09:34:47 nerdyRAID cnid_metad[17126]: Multiple attempts to start CNID db daemon for "/mnt/user/macUP" failed, wiping the slate clean...
Jan 6 09:34:47 nerdyRAID cnid_dbd[17126]: error opening DB environment after recovery: No such device
Jan 6 09:34:47 nerdyRAID cnid_dbd[17126]: Failed to open CNID database for volume "macUP"
Jan 6 09:34:47 nerdyRAID cnid_dbd[17126]: Recreated CNID BerkeleyDB databases of volume "macUP"
Jan 6 09:34:47 nerdyRAID cnid_dbd[17126]: error opening DB environment after recovery: No such device
Jan 6 09:34:47 nerdyRAID cnid_dbd[17126]: open_db() failed: No such device
Jan 6 09:34:47 nerdyRAID cnid_dbd[17126]: reinit_db() failed: No such device
Jan 6 09:34:47 nerdyRAID afpd[17065]: read: Connection reset by peer
Jan 6 09:34:48 nerdyRAID cnid_dbd[17132]: error opening DB environment after recovery: No such device
Jan 6 09:34:48 nerdyRAID cnid_dbd[17132]: Failed to open CNID database for volume "macUP"
Jan 6 09:34:48 nerdyRAID cnid_dbd[17132]: Recreated CNID BerkeleyDB databases of volume "macUP"
Jan 6 09:34:48 nerdyRAID cnid_dbd[17132]: error opening DB environment after recovery: No such device
Jan 6 09:34:48 nerdyRAID cnid_dbd[17132]: open_db() failed: No such device
Jan 6 09:34:48 nerdyRAID cnid_dbd[17132]: reinit_db() failed: No such device
Jan 6 09:34:48 nerdyRAID afpd[17065]: read: Connection reset by peer
Jan 6 09:34:49 nerdyRAID afpd[17065]: read: Connection reset by peer
Jan 6 09:34:50 nerdyRAID afpd[17065]: read: Connection reset by peer
Jan 6 09:34:51 nerdyRAID afpd[17065]: read: Connection reset by peer
Jan 6 09:34:52 nerdyRAID afpd[17065]: read: Connection reset by peer
Jan 6 09:34:53 nerdyRAID afpd[17065]: read: Connection reset by peer
Jan 6 09:34:54 nerdyRAID afpd[17065]: read: Connection reset by peer
Jan 6 09:34:55 nerdyRAID afpd[17065]: read: Connection reset by peer
Jan 6 09:34:56 nerdyRAID afpd[17065]: read: Connection reset by peer
Jan 6 09:34:57 nerdyRAID afpd[17065]: read: Connection reset by peer

 

Searching forum, and, google -- having difficulty finding a viable solution (probably just me)\

 

Looking for suggestions

 

What users are setup to access the share and what are the share settings? Typically when you try to connect TM to a share it will prompt you for credentials to connect to that share, are you getting prompted for credentials?

Set to Secure

 

User Jnerdy (access: read/write)

Credentials prompted

 

Then time outs and says server can not be found

Try an SMB share just to see what the results are.

You have Netatalk database corruption on that share. Make sure it isn't AFP mounted and delete the .AppleDB folder in the root of the share:

 

rm -rf /mnt/user/macUP/.AppleDB

 

The next time you mount it using the AFP protocol the folder and database will be recreated. I moved my Netatalk databases off the array and onto my cache pool and I find AFP mounts much more reliable and responsive.

 

https://lime-technology.com/forum/index.php?topic=51632.msg520908#msg520908

 

My /mnt/user/system is a cache-only share (and therefore is equivalent to /mnt/cache/system) and if you were to do the same as me your Netatalk database would be relocated to /mnt/user/system/AppleDB/macUP/.AppleDB (equivalent to /mnt/cache/system/AppleDB/macUP/.AppleDB).

 

Thanks John_M

 

Counter... would it be better to just create a sparsebundle and point macOS to the unRAID share over SMB...and cut AFP out altogether?

 

When a network volume is used as Time Machine's backup destination a sparse bundle is used anyway but at the moment only AFP provides the support needed for Time Machine to work. The situation is complicated by the fact Time Machine first has to mount the remote volume, then it has to mount the sparse bundle disk image. These two may have different permissions and ownerships and it's easy to get things wrong.

 

Sierra allows SMB to be used instead of AFP but on condition that it provides a full set of extensions. Some but not all of those extensions are provided in unRAID 6.3.0-rc6 (you don't say which version of unRAID you are using) so, while SMB support for Time Machine is close, it isn't quite there yet. For more information see the earlier entries in that thread I referenced above.

 

For general Time Machine information see here.

 

Thanks John_M

 

I previously had a HFS+ sparsebundle mounted in an WD mybook live, and was going to try and use the same process.

 

 

my /mnt/user/system is also cache only, so you suggest  dbpath  change to /mnt/user/system?

 

sorry:  6.2.4

 

 

You can set the dbpath to /mnt/user/system if you want to. Using the example above, your macUP share's .AppleDB will be migrated to /mnt/user/system/macUP/.AppleDB

 

I have other subfolders within my system share and wanted to keep the Netatalk databases together, rather than cluttering up its root so I set my dbpath to the path I indicated. What ever works for you, really.

 

The database can grow quite large and consists of a number of individual files within the .AppleDB folder and in the case of a user share can end up split across several physical disks - that was what prompted me to move it to the cache.

 

Time Machine backup destinations have to be HFS+ formatted (preferably with the journalling option enabled), hence the need to use a disk image when the physical destination is across a network. A sparse bundle is preferred because of its granularity (being composed of many small (8 MB) files). AFP is currently still needed for it to work reliably.

 

Understood.

 

Its a shame, since HFS+ as a FS is getting long in the tooth.

 

Reading suggests that AFP could be deprecated over the next few point updates (in favor of SMB)...though that is speculation.  The announcement of the new AFS will hopefully bring a robust set of conditions for network based disk imaging and incremental backup.

 

I will try your suggestions later this evening.

 

I will also try your folder structure...

 

/mnt/user/system/AppleDB

 

the folder  /macUP will then be recreated with .AppleDB nested within it, correct? 

 

I used

rm -rf 

to delete .AppleDB and MC to delete the share. 

 

 

I seem to remember Linus Torvalds once remarking that HFS+ is the worst file system ever...

 

AFP won't go way until SMB provides the necessary functionality. See Tom's post (limetech) in the thread I gave in my first reply.

 

Regarding AFS, I won't be holding my breath...

 

Yes, the sub-folder named for the share and the .AppleDB folder within it and the database files within that are all created automatically the first time you AFP-mount the share. It may take a little while for the Finder to open it the first time, especially if you don't first remove the old database and the system has to relocate it. But you will hopefully find, as I have, that subsequent mounts are smoother and more reliable.

 

Unfortunately, when I mount the share via GUI the log explodes with same error.

 

I did pull this form errlog in .AppleDB

 

Finding last valid log LSN: file: 1 offset 28
mmap: No such device
mmap: No such device
mmap: No such device
Finding last valid log LSN: file: 1 offset 28
mmap: No such device
mmap: No such device
mmap: No such device
Finding last valid log LSN: file: 1 offset 28
mmap: No such device
mmap: No such device
mmap: No such device
Finding last valid log LSN: file: 1 offset 28
mmap: No such device
mmap: No such device

 

I am going to delete and try again

 

Add Share: macUP

Highwater

500MB min free

Split as needed

Disk4 assigned

AFP

Export Yes

Time Machine

512000MB limit

dbpath /mnt/user/system/AppleDBJan 7 02:49:30

 

nerdyRAID emhttp: shcmd (67308): mkdir '/mnt/user/macUP' |& logger
Jan 7 02:49:30 nerdyRAID emhttp: shcmd (67309): chmod 0777 '/mnt/user/macUP'
Jan 7 02:49:30 nerdyRAID emhttp: shcmd (67310): chown 'nobody':'users' '/mnt/user/macUP'
Jan 7 02:50:55 nerdyRAID emhttp: shcmd (67380): mkdir -p /mnt/user/system/AppleDB
Jan 7 02:50:55 nerdyRAID avahi-daemon[19967]: Files changed, reloading.
Jan 7 02:50:55 nerdyRAID avahi-daemon[19967]: Service group file /services/afp.service changed, reloading.
Jan 7 02:50:56 nerdyRAID avahi-daemon[19967]: Service "nerdyRAID-AFP" (/services/afp.service) successfully established.

 

And once  I enter why credentials

 

Jan 7 03:04:38 nerdyRAID cnid_dbd[8125]: error opening DB environment after recovery: No such device
Jan 7 03:04:38 nerdyRAID cnid_dbd[8125]: Failed to open CNID database for volume "macUP"
Jan 7 03:04:38 nerdyRAID cnid_dbd[8125]: Recreated CNID BerkeleyDB databases of volume "macUP"
Jan 7 03:04:38 nerdyRAID cnid_dbd[8125]: error opening DB environment after recovery: No such device
Jan 7 03:04:38 nerdyRAID cnid_dbd[8125]: open_db() failed: No such device
Jan 7 03:04:38 nerdyRAID cnid_dbd[8125]: reinit_db() failed: No such device
Jan 7 03:04:38 nerdyRAID afpd[8059]: read: Connection reset by peer
Jan 7 03:04:39 nerdyRAID cnid_dbd[8135]: error opening DB environment after recovery: No such device
Jan 7 03:04:39 nerdyRAID cnid_dbd[8135]: Failed to open CNID database for volume "macUP"
Jan 7 03:04:39 nerdyRAID cnid_dbd[8135]: Recreated CNID BerkeleyDB databases of volume "macUP"
Jan 7 03:04:39 nerdyRAID cnid_dbd[8135]: error opening DB environment after recovery: No such device
Jan 7 03:04:39 nerdyRAID cnid_dbd[8135]: open_db() failed: No such device
Jan 7 03:04:39 nerdyRAID cnid_dbd[8135]: reinit_db() failed: No such device
Jan 7 03:04:39 nerdyRAID afpd[8059]: read: Connection reset by peer
Jan 7 03:04:40 nerdyRAID cnid_metad[8141]: Multiple attempts to start CNID db daemon for "/mnt/user/macUP" failed, wiping the slate clean...
Jan 7 03:04:40 nerdyRAID cnid_dbd[8141]: error opening DB environment after recovery: No such device
Jan 7 03:04:40 nerdyRAID cnid_dbd[8141]: Failed to open CNID database for volume "macUP"
Jan 7 03:04:40 nerdyRAID cnid_dbd[8141]: Recreated CNID BerkeleyDB databases of volume "macUP"
Jan 7 03:04:40 nerdyRAID cnid_dbd[8141]: error opening DB environment after recovery: No such device
Jan 7 03:04:40 nerdyRAID cnid_dbd[8141]: open_db() failed: No such device
Jan 7 03:04:40 nerdyRAID cnid_dbd[8141]: reinit_db() failed: No such device
Jan 7 03:04:40 nerdyRAID afpd[8059]: read: Connection reset by peer
Jan 7 03:04:41 nerdyRAID cnid_dbd[8150]: error opening DB environment after recovery: No such device
Jan 7 03:04:41 nerdyRAID cnid_dbd[8150]: Failed to open CNID database for volume "macUP"
Jan 7 03:04:41 nerdyRAID cnid_dbd[8150]: Recreated CNID BerkeleyDB databases of volume "macUP"
Jan 7 03:04:41 nerdyRAID cnid_dbd[8150]: error opening DB environment after recovery: No such device
Jan 7 03:04:41 nerdyRAID cnid_dbd[8150]: open_db() failed: No such device
Jan 7 03:04:41 nerdyRAID cnid_dbd[8150]: reinit_db() failed: No such device
Jan 7 03:04:41 nerdyRAID afpd[8059]: read: Connection reset by peer
Jan 7 03:04:42 nerdyRAID afpd[8059]: read: Connection reset by peer
Jan 7 03:04:43 nerdyRAID afpd[8059]: read: Connection reset by peer
Jan 7 03:04:44 nerdyRAID afpd[8059]: read: Connection reset by peer
Jan 7 03:04:45 nerdyRAID afpd[8059]: read: Connection reset by peer
Jan 7 03:04:46 nerdyRAID afpd[8059]: read: Connection reset by peer
Jan 7 03:04:47 nerdyRAID afpd[8059]: read: Connection reset by peer
Jan 7 03:04:47 nerdyRAID afpd[8059]: transmit: Request to dbd daemon (volume macUP) timed out.
Jan 7 03:04:47 nerdyRAID afpd[8059]: afp_openvol(/mnt/user/macUP): Fatal error: Unable to get stamp value from CNID backend
Jan 7 03:04:47 nerdyRAID afpd[8059]: dsi_stream_read: len:0, unexpected EOF

I've been fighting timemachine for over a year now, sometimes it works but mostly not. Usually only after i try a 1000 times to select the disk using one of many credentials, and then suddenly it finds the unraid disk and starts backup. Lately it was backing up for 90% and then stopped with some *&^&*&* message ' disk in use' or some other f*cked up message that could not be fixed in any way. Sometimes i even had to totally delete the share with the sparsebundle and could only backup an initial, new backup, no incremental ones.

 

I ended up not using that stupid timemachine nonsense and am using Carbon Copy Cloner instead. Nicely scheduled once a week, sends an email when done, creates bootable backups if you wish. Works like a charm.

I've been fighting timemachine for over a year now, sometimes it works but mostly not. Usually only after i try a 1000 times to select the disk using one of many credentials, and then suddenly it finds the unraid disk and starts backup. Lately it was backing up for 90% and then stopped with some *&^&*&* message ' disk in use' or some other f*cked up message that could not be fixed in any way. Sometimes i even had to totally delete the share with the sparsebundle and could only backup an initial, new backup, no incremental ones.

 

I ended up not using that stupid timemachine nonsense and am using Carbon Copy Cloner instead. Nicely scheduled once a week, sends an email when done, creates bootable backups if you wish. Works like a charm.

 

Wow.

 

I would actually prefer a disk imaging tool. Thank you for the suggestion, I will definitely look into it.  Macrium is my goto tool in the winters

 

Time Machine using a USB-connected hard disk that's dedicated to the task is much more reliable than using a file server as a destination. Carbon Copy Cloner is good, as is SuperDuper!