Timemachine: The network backup disk does not support the required AFP feature

On my Backup Box which runs since months there is a new problem with one client. Neither the OMV machine (1.19) nor the OS X (10.9.x) client got changed, no major upgrades.

Sometimes, and only sometimes the TS backup fails with the following error:

The network backup disk does not support the required AFP feature

In the afp logs there are just the usual (dsi_stream) errors (which are a feature of the old netatalk 2.2 afaik):

May 24 14:47:27 bus afpd[28306]: AFP3.3 Login by nk
May 24 14:47:27 bus afpd[28306]: afp_disconnect: trying primary reconnect
May 24 14:47:27 bus afpd[2700]: Reconnect: transfering session to child[3656]
May 24 14:47:27 bus afpd[2700]: Reconnect: killing new session child[28306] after transfer
May 24 14:47:27 bus afpd[3656]: afp_dsi_transfer_session: succesfull primary reconnect
May 24 14:47:27 bus afpd[3656]: dsi_stream_read: len:0, unexpected EOF
May 24 14:47:27 bus afpd[3656]: dsi_stream_read: len:0, unexpected EOF
May 24 14:47:27 bus afpd[3656]: dsi_disconnect: entering disconnected state
May 24 14:47:28 bus afpd[28241]: afp_disconnect: primary reconnect succeeded
May 24 14:47:28 bus afpd[28312]: AFP3.3 Login by nk
May 24 14:47:28 bus afpd[28312]: afp_disconnect: trying primary reconnect
May 24 14:47:28 bus afpd[2700]: Reconnect: transfering session to child[3656]
May 24 14:47:28 bus afpd[2700]: Reconnect: killing new session child[28312] after transfer
May 24 14:47:28 bus afpd[3656]: afp_dsi_transfer_session: succesfull primary reconnect
May 24 14:47:28 bus afpd[28247]: afp_disconnect: primary reconnect succeeded
May 24 14:47:28 bus afpd[28252]: afp_disconnect: primary reconnect succeeded
May 24 14:47:28 bus afpd[28258]: afp_disconnect: primary reconnect succeeded
May 24 14:47:28 bus afpd[28264]: afp_disconnect: primary reconnect succeeded
May 24 14:47:28 bus afpd[28270]: afp_disconnect: primary reconnect succeeded
May 24 14:47:29 bus afpd[28276]: afp_disconnect: primary reconnect succeeded
May 24 14:47:29 bus afpd[28283]: afp_disconnect: primary reconnect succeeded
May 24 14:47:29 bus afpd[28289]: afp_disconnect: primary reconnect succeeded
May 24 14:47:29 bus afpd[28294]: afp_disconnect: primary reconnect succeeded
May 24 14:47:29 bus afpd[28300]: afp_disconnect: primary reconnect succeeded
May 24 14:47:29 bus afpd[28306]: afp_disconnect: primary reconnect succeeded
May 24 14:47:30 bus afpd[28312]: afp_disconnect: primary reconnect succeeded
May 24 14:47:51 bus afpd[3656]: AFP logout by nk
May 24 14:47:51 bus afpd[3656]: AFP statistics: 460158.42 KB read, 266990.92 KB written
May 24 14:47:51 bus afpd[3656]: done

Its on Ethernet, so no wifi issues.

Any ideas?

Ok, will try. Thanks.

Thinking about dumping the AFP Gui and installing Netatalk 3 manually...

Btw. how can i block the netatalk plugin, so its not selected by (user)error. Any way to do this?

(Dont want to override my manual netatalk 3 installation)

Thanks for the info. Conversation was before 3.17?

Asking because 3.17 has already the new compile flags for debian (systemd/sysv).

Did compile the 3.17 and the 3.18dev (git) without problems...

P.S. VMs are for pussies .. compile on the production server ... better then latenight crime tv

Hmm, this was around the 3.17 release. Not sure if the mentionend batches are in 3.17 already...

Anyways, tried with the 3.17 release, compiled fine, init scripts installed, running, works fine.

Blackmagic Speed test: 108MB write, 110MB read ... so far so good

So, timemachine backups where running flawless. For now quite happy.

Didnt configure Netatalks 3 Spotlight, will do maybe later.

Btw. i think the build options from the maintainer are not complete, for omv there should be acl support and some more compiled in

Have to look in the doku, but i enabled it in configure script.

Anyways, one small problem remains. Permission problem for the size check of TM sparsebundle file. Have to look into that.

But ... got a failed TM backup error again .. bummer

Didnt find much about it here.

But ... i did a permission reset, deleted the content of the backup folders. After this did backup from two clients at the same time, worked good. Hope it stays like that.

SubZero: in my version most options are in:

afpd 3.1.7 - Apple Filing Protocol (AFP) daemon of Netatalk

This program is free software; you can redistribute it and/or modify it under
the terms of the GNU General Public License as published by the Free Software
Foundation; either version 2 of the License, or (at your option) any later
version. Please see the file COPYING for further information and details.

afpd has been compiled with support for these features:

AFP versions: 2.2 3.0 3.1 3.2 3.3 3.4
CNID backends: dbd last tdb mysql
Zeroconf support: Avahi
TCP wrappers support: Yes
Quota support: Yes
Admin group support: Yes
Valid shell checks: Yes
cracklib support: Yes
EA support: ad | sys
ACL support: Yes
LDAP support: Yes
D-Bus support: Yes
Spotlight support: Yes
DTrace probes: Yes

afp.conf: /etc/netatalk/afp.conf
extmap.conf: /etc/netatalk/extmap.conf
state directory: /var/lib/netatalk/
afp_signature.conf: /var/lib/netatalk/afp_signature.conf
afp_voluuid.conf: /var/lib/netatalk/afp_voluuid.conf
UAM search path: /usr/lib/netatalk//
Server messages path: /var/lib/netatalk/msg/

---

@wastl: My TM stuff was running without major problems for months. But a few days ago troubles started. So i turned to netatalk 3.x. And yes, its still running on my machines.

Btw seems that the dsi_stream errors are gone for now with 3.17 (which afaik is a netatalk 2.2 bug)

Update:

Backup during the night from 2 clients at same time ok
Backup from 3rd client during the day while 2 clients doing incremental backups ok

For 2 clients i used the same share and login name, also without problems.

No error msgs in logs.

Nice transferrates (tm is always slow) up to 750mb/s on omv network interface. Avg around 400-500 (from a new mbp with ssd)

What wonders me ... why is TM searching for the backup discs always that long on network shares? Had this on all backup servers, no matter omv, freenas, netatalk 2 or 3. What is osx doing that long???

Guys, new question where i got stuck (maybe i should go to bed....)

Im playing around with a special configuration (on stoneburner):

I have an extra LXC container with netatalk3 running, in addition to the netatalk3 in the root system.

Reason: this LXC container is connected to another ethernet card to provide a second (isolated) backup system for a different lan segment (i know the lxc isolation is not perfect, but i can live with that for now)

So far so good. What im wondering ... how is the user authentication (shared folder access rights) handled between omv and netatalk? My netatalk3 on the root system works fine. Im thinking about connecting the netatalk3 lxc instance to the root system for user authentication, so i can manage the users still from omv gui.