AFP works - sometimes (!?) - need help to get it stable

[tak21]

solved - AFP works - sometimes (!?) - need help to get it stable

My freenas server is used to serve the MACs at home as tm-backup and fileserver.
But there are times, when I have no access to the server. As an example today, right after reboot. No access (11-10 21:36).
And then it starts to work again (12-10 01:06). And beside from what I just described the system tends to stop on a regular base (stop= no access at all via any protocol).
What can I do to get a stable system? (And please let me know, if more info’s are required!)

/var/log/messages

View attachment messages.txt

System Information
Hostname minisrv.local
Build FreeNAS-8.3.0-RC1-x64 (r12617)
Platform Intel(R) Atom(TM) CPU N2800 @ 1.86GHz
Memory 8159MB
System Time Fri Oct 12 01:17:34 CEST 2012
Uptime 1:17AM up 4:35, 1 user
Load Average 0.09, 0.10, 0.08
Connected through minisrv

• user "share" has full access to all three zfs datasets used for sharing (1. timemachine, 2. data, 3. sw-mac)
• user "share" is used as guest access user in the afp service set up
• all shares were set up with the following parameters (except tm - this share has the discovery mode tm)

Code:

Name [ *** ]
Share Comment [ *** - AFP share ]
Path [ /mnt/tank/*** ]
Allow List [ share ]
Read-write Access [ share ]
Disk Discovery [ x ]
Disk discovery mode [ Default ]
Cache CNID [ x ]
Enable .AppleDouble [ x ]
Default file permission
		ow	gr	oth
read		 x	 x	 x
write		 x	 x
execute	 x	 x	 x

Default directory permission
		ow	gr	oth
read		 x	 x	 x
write		 x	 x
execute	 x	 x

 

[JaimieV]

My own primary system is set up much the same as yours (one less dataset), and gives no such problems.

If the whole system "stops", then this isn't an AFP problem but a system one. I'd consider this as the priority problem! Do you have a messages.log of an example stop?
I see this one complains that / and /data were not properly dismounted. Did the NAS crash and need restarting, or does it normally pause and recover all by itself? Or did you happen to power it down manually to get it to come back? Not counting the AFP freeze, which we can deal with later.

Have you checked the SMART status of the drives? Bad block re-reads are one of the things that could cause a file-sharing freeze, but is unlikely to cause a system wide freeze. Worth looking anyway.

 

[tak21]

For the complain: I did pull the trigger, because it did not work anymore.

What I did change by now:
- reloaded and reinstalled FreeNAS-8.3.0-RC1-x64 (r12617)
- from the AFP settings - per share - removed the [x] for "AFP3 Unix Privs"
It now works stable. I am backin up tree macs and have full access to my data (and do a reguar backup of the little server because there is still a little misstrust left...)