AFP-Shares disapear and comes back again

Hessi · 25. Februar 2015

Hi Folks,

since the last Updates in late january, I've got no status emails from the NAS. I dind't realised it until the mchine gets very slow transferrates.
Therefore I had a look at the updates on the webend. But the system can't connect to any external service, because of the gateway at the manual network settings are empty now!
Re-entered it an run several updates.

Now, I connect to the shares and while I am working on it, it disapears under my mouse cursor!
Then... suddenly the apear again, sometimes empty, sometimes with less folders...

The transfer rates are absolutly slow, too.

Ah! If I activate SMB, sometimes (without any connection) it takes 1 Core of the CPU with 100%. But I sometime, I cannot see the service in top or at the webend, sometimes it is SMB what driving nuts.

I have remove (purged) openmediavault.netatalk and netatalk and installed it after a reboot.
Also remove all other service discovery instead of AFP.

Got many errory, have a look at the logs.

Code

Feb 25 21:02:11 NAS afpd[13623]: ===============================================================
Feb 25 21:02:11 NAS afpd[13623]: INTERNAL ERROR: Signal 11 in pid 13623 (2.2.2)
Feb 25 21:02:11 NAS afpd[13623]: ===============================================================
Feb 25 21:02:11 NAS afpd[13623]: BACKTRACE: 12 stack frames:
Feb 25 21:02:11 NAS afpd[13623]:  #0 /usr/sbin/afpd(netatalk_panic+0x1c) [0x7f73204e0e4c]
Feb 25 21:02:11 NAS afpd[13623]:  #1 /usr/sbin/afpd(+0x55f4c) [0x7f73204e0f4c]
Feb 25 21:02:11 NAS afpd[13623]:  #2 /lib/x86_64-linux-gnu/libc.so.6(+0x321e0) [0x7f731e9db1e0]
Feb 25 21:02:11 NAS afpd[13623]:  #3 /usr/sbin/afpd(bdestroy+0xe) [0x7f73204da99e]
Feb 25 21:02:11 NAS afpd[13623]:  #4 /usr/sbin/afpd(dir_add+0x2bb) [0x7f73204ae18b]
Feb 25 21:02:11 NAS afpd[13623]:  #5 /usr/sbin/afpd(cname+0x7e7) [0x7f73204af967]
Feb 25 21:02:11 NAS afpd[13623]:  #6 /usr/sbin/afpd(afp_getfildirparams+0xd6) [0x7f73204b9c06]
Feb 25 21:02:11 NAS afpd[13623]:  #7 /usr/sbin/afpd(afp_over_dsi+0x50c) [0x7f73204a294c]
Feb 25 21:02:11 NAS afpd[13623]:  #8 /usr/sbin/afpd(+0x16606) [0x7f73204a1606]
Feb 25 21:02:11 NAS afpd[13623]:  #9 /usr/sbin/afpd(main+0x77e) [0x7f732049ebce]
Feb 25 21:02:11 NAS afpd[13623]:  #10 /lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xfd) [0x7f731e9c7ead]
Feb 25 21:02:11 NAS afpd[13623]:  #11 /usr/sbin/afpd(+0x1427d) [0x7f732049f27d]
Feb 25 21:02:11 NAS afpd[13630]: AFP3.3 Login by michael
Feb 25 21:02:11 NAS afpd[13630]: afp_disconnect: trying primary reconnect
Feb 25 21:02:11 NAS afpd[10164]: Reconnect: no child[13623]
Feb 25 21:02:16 NAS afpd[13630]: afp_disconnect: primary reconnect failed
Feb 25 21:02:17 NAS afpd[13630]: getfd: getsockopt says: Connection refused
Feb 25 21:02:17 NAS afpd[13630]: getfd: getsockopt says: Connection refused
Feb 25 21:02:17 NAS afpd[13630]: tsock_getfd: no suitable network config from CNID server (localhost:4700): Connection refused
Feb 25 21:02:17 NAS afpd[13630]: transmit: connection refused (db_dir /media/fa12bb93-c569-4b55-b80b-7cdd186ddede/TimeMachine)
Feb 25 21:02:17 NAS afpd[13630]: Reopen volume /media/fa12bb93-c569-4b55-b80b-7cdd186ddede/TimeMachine using in memory temporary CNID DB.
Feb 25 21:02:34 NAS afpd[13630]: getfd: getsockopt says: Connection refused
Feb 25 21:02:34 NAS afpd[13630]: getfd: getsockopt says: Connection refused
Feb 25 21:02:34 NAS afpd[13630]: tsock_getfd: no suitable network config from CNID server (localhost:4700): Connection refused
Feb 25 21:02:34 NAS afpd[13630]: transmit: connection refused (db_dir /media/a17632c2-23b3-4d60-a7ab-ca18b5b54cc7)
Feb 25 21:02:34 NAS afpd[13630]: getfd: getsockopt says: Connection refused
Feb 25 21:02:34 NAS afpd[13630]: getfd: getsockopt says: Connection refused
Feb 25 21:02:34 NAS afpd[13630]: tsock_getfd: no suitable network config from CNID server (localhost:4700): Connection refused
Feb 25 21:02:34 NAS afpd[13630]: transmit: connection refused (db_dir /home/michael)
Feb 25 21:02:34 NAS afpd[13630]: ===============================================================
Feb 25 21:02:34 NAS afpd[13630]: INTERNAL ERROR: Signal 11 in pid 13630 (2.2.2)
Feb 25 21:02:34 NAS afpd[13630]: ===============================================================
Feb 25 21:02:34 NAS afpd[13630]: BACKTRACE: 11 stack frames:
Feb 25 21:02:34 NAS afpd[13630]:  #0 /usr/sbin/afpd(netatalk_panic+0x1c) [0x7f73204e0e4c]
Feb 25 21:02:34 NAS afpd[13630]:  #1 /usr/sbin/afpd(+0x55f4c) [0x7f73204e0f4c]
Feb 25 21:02:34 NAS afpd[13630]:  #2 /lib/x86_64-linux-gnu/libc.so.6(+0x321e0) [0x7f731e9db1e0]
Feb 25 21:02:34 NAS afpd[13630]:  #3 /usr/sbin/afpd(bdestroy+0xe) [0x7f73204da99e]
Feb 25 21:02:34 NAS afpd[13630]:  #4 /usr/sbin/afpd(dir_add+0x2bb) [0x7f73204ae18b]
Feb 25 21:02:34 NAS afpd[13630]:  #5 /usr/sbin/afpd(+0x280c8) [0x7f73204b30c8]
Feb 25 21:02:34 NAS afpd[13630]:  #6 /usr/sbin/afpd(afp_over_dsi+0x50c) [0x7f73204a294c]
Feb 25 21:02:34 NAS afpd[13630]:  #7 /usr/sbin/afpd(+0x16606) [0x7f73204a1606]
Feb 25 21:02:34 NAS afpd[13630]:  #8 /usr/sbin/afpd(main+0x77e) [0x7f732049ebce]
Feb 25 21:02:34 NAS afpd[13630]:  #9 /lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xfd) [0x7f731e9c7ead]
Feb 25 21:02:34 NAS afpd[13630]:  #10 /usr/sbin/afpd(+0x1427d) [0x7f732049f27d]
Feb 25 21:02:34 NAS afpd[13672]: AFP3.3 Login by michael
Feb 25 21:02:34 NAS afpd[13672]: afp_disconnect: trying primary reconnect
Feb 25 21:02:34 NAS afpd[10164]: Reconnect: no child[13630]

Alles anzeigen

Systeminformations: ubt0T5nm

TIA
Regards
Michael

subzero79 · 25. Februar 2015

Are you still using and SSD for system drive?

Hessi · 25. Februar 2015

Yes.

subzero79 · 26. Februar 2015

I was looking at your log and I didnt see the discard (trim) option in fstab.
Do you know if your drive supports trim?

Hessi · 26. Februar 2015

"Discard" means "TRIM enabled"?
In thsi case, TRIM is enabled:
UUID=a770f0b1-0f3b-44c1-8b55-2fb67cb48781 / ext4 noatime,nodiratime,discard,errors=remount-ro 0 1

subzero79 · 26. Februar 2015

It doesn't mean is supported. It means it will trim everytime at boot.

Whats the ssd brand model?

Hessi · 26. Februar 2015

Yes, it does:
http://download.intel.com/supp…Product_specification.pdf

subzero79 · 26. Februar 2015

Well the stack trace is software failure for afpd. For that we cannot do much here. Also if you notice netatalk version is very outdated in debian, I have updated to 2.2.5 from Jessie just to use symlinks. I don't use time machine but I can give it a try later to see I come with the same or similar, what's you osx version?
Just a search in Google lead me to this page in Ubuntu from 2012. You can check there if one of the solutions there work for you.
https://bugs.launchpad.net/ubu…urce/netatalk/+bug/810732

Hessi · 26. Februar 2015

We have one 10.9 and two 10.10 machines.

Edit:
Can you tell me how to find out, what is going on?
Do you see the difference? This is typical behaivior for round about a month.

I followed these instructions:
http://netatalk.sourceforge.ne….7_on_Ubuntu_14.10_Utopic
Finaly: apt-get install openmediavault-netatalk

The Performance is ca. 30% less than all my benchmarks (Black Magic Speed Test) after the fresh installation of OMV.
Get only 70/80 megs/sec.

But in RL, I need actual 3h for 73 GB!

Edit: It seems to be stable since several hours. Also TM is working fine: CLOSED!

subzero79 · 26. Februar 2015

Netatalk 3 uses a whole new syntax. I have real doubts that the plugin is gonna work there, unless you configure it manually in CLI, then I guess you don't need the plugin.

Also there is a netatalk3 that was tried to pushed to Jessie and couldn't make it here https://github.com/adiknoth/netatalk-debian/

You need just to revert a specific commit thats specific to Jessie

ryecoaaron · 26. Februar 2015

I just noticed that there is no netatalk package in jessie - link

Hessi · 26. Februar 2015

No, the 3.1.7 Netatalk worked with openmediavault.netatalk-plugin as it should be.
I configured it via webend and it seems to be still stable.
I guess, the syntax for the config files are compatible.

subzero79 · 26. Februar 2015

they failed to get before the freeze. I've sent you the email I've discussed with the maintainer.
The reason is in some mail list for debian netatalk.

subzero79 · 27. Februar 2015

I a still have doubts that is working.
First is the netatalk 2.2.5 being removed?

How did you install netatalk 3.1, make install? Or you converted the built into a deb package?

If the plugin is working I guess is because is still configuring 2.2.5, you probably have two different afpd binaries installed at different path. Just check with "afpd - V" at /usr/bin and at /usr/local/bin

Inspect the init script and check which binary is using to start the daemon.

Hessi · 27. Februar 2015

Of course, I have removed netatalk first. Then I installed the new version with make.

subzero79 · 27. Februar 2015

You shouldn't use make install. You should use checkinstall. This generates a deb package that's easily removed with apt-get.
Also at some point the plugin will try to install the old version. Watch that.

Hessi · 27. Februar 2015

I installed checkinstall and created the package, and still get this versioninfo:
"afpd 3.1.7 - Apple Filing Protocol (AFP) daemon of Netatalk"

subzero79 · 27. Februar 2015

Ok. That sounds better, however the paths used by that Ubuntu version are not the ones in debian, but if it works for you then I guess is alright.
Can you post here the contents of this file /usr/local/etc/afp.conf

Hessi · 27. Februar 2015

I'm driving nuts! My shares are gone away! NAS isn't avaible via afp anymore! Last working TM-Backup is three hours ago?
The system doesn't fell asleep all the time, cause my doughter is ill and watching movies all the time with the Win 8.1-HPTC via smb-shares.
WTF?
I guess, you're right and it wasn't a good idea to install netatalk 3.7.1
Okay, I will give it a second try this evening... or so.

subzero79 · 27. Februar 2015

You can use netatalk 3. Just don't use the plugin and configure shares manually.

Jetzt mitmachen!