atalk hangs and results in a variety of issues

[cdonner]

TS-495U, Firmware 3.6.1 Build 0302T. 4x2TB disks in a Raid 5 array. One user uses TimeMachine for backup.
This morning users reported problems accessing the NAS via SMB, SSH, FTP ...
A log showed an "out of disk space" condition as of last Friday, but today there are over 1.6TB available.
The admin site was accessible, but certain pages timed out. Only a hard reset would get the NAS to reboot. The problems persisted after a reboot.

A look at the console identified the atalk demon as a potential culprit. The line that starts it was the last one to appear before the screen froze.
I disabled TimeMachine and the AFP service and everything returned to normal.

The TimeMachine user wants this fixed. I am going to try to remove all existing backups on the NAS and start with a clean slate. Did anybody else see this issue, and are there any other suggestions?

[cdonner]

I spent an hour with QNAP support last week. Apparently the issue is no unknown to them, and the person made some configuration file changes and rebooted several times. I was warned that I would lose network shares and permissions in the process, but asked to proceed anyways.

When we were done and I was off the phone again, I rejoined the domain and enabled TimeMachine. I immediately was no longer able to access the Windows Networking or the Systems Information pages and the NAS could not be reached via SSH. This situation persists until I turn TimeMachine off, then the services can be reached again. I can repeat this cycle with the same results - when TimeMachine & AFP run, the NAS goes partially out of service but returns to normal when TimeMachine is turned off.
The bottom line is that I am still looking for a solution, because I need TimeMachine.

[ghilteras]

exactly the same issue.

As soon as I enable time machine Atalk hangs, ssh and samba are stuck.

Need a fix ASAP

[ghilteras]

any news about it? is a fix coming?

[cdonner]

I forgot to update this thread - I spent time with support and they determined that it only happens in conjunction with RR load balancing. Once I changed the load balancing configuration to hot stand-by, the problems went away. Still, this isn't an actual fix, but I'd imagine that it is serious enough that Qnap made it a priority.

[P3R]

I spent time with support and they determined that it only happens in conjunction with RR load balancing.

So did you have a manageable switch configured for a static trunk on the ports the Qnap connects to? If not, it was a misconfigured network that caused your problems and no problem Qnap can fix.

[ghilteras]

if a switch does not support balance rr it simply does not work, how can you say it's not qnap's fault if it crashes enabling atalk? it's clearly a qnap issue.

moreover we cannot disable rr just to make time machine work, we NEED a fix for it, regardless of swtiches being compliant or not

[doktornotor]

we NEED a fix for it, regardless of swtiches being compliant or not

LOLz? So you want QNAP to fix something caused by using settings/features that your network HW does not support? Good luck with that.

P.S.

I would suggest to only enable features
- you are able to handle yourself (i.e., you understand how actually the feature works and what are the requirements to utilize);and
- which you are 100% sure your HW can handle as well

Not following the above advice will not only not improve the network throughput and performance but will - on the contrary - lead to a total discruption of service.

[cdonner]

doktormotor, remember that this worked properly for years until the Firmware 3.6.1 Build 0302T. You can argue that the firmware was broken and that it should not have allowed a misconfigured scenario, but I don't think users really care about that.

[doktornotor]

Yeah, makes sense. Completelely misconfigured trunking with switches non supporting any of that => QNAP issue. Oh well. No point in arguing here, I guess, I have better things to do with my time.

P.S. Your "worked properly" probably meaning the load ballancing did not work at all. So now that it works, it bombs out due to being used on a HW that does not support it.

[ghilteras]

we NEED a fix for it, regardless of swtiches being compliant or not

LOLz? So you want QNAP to fix something caused by using settings/features that your network HW does not support? Good luck with that.

P.S.

I would suggest to only enable features
- you are able to handle yourself (i.e., you understand how actually the feature works and what are the requirements to utilize);and
- which you are 100% sure your HW can handle as well

Not following the above advice will not only not improve the network throughput and performance but will - on the contrary - lead to a total discruption of service.

is it a joke? qnap has a bug. it hangs when atalk and rr bonding are enabled together, I payed for the machine so I'm asking for a fix, no trolling.


so, back to the original question, is anyone on the staff willing to give an answer? is this bug being treated? can we have an ETA?

[doktornotor]

is it a joke? qnap has a bug.

Once again, you have a PEBKAC bug enabling features unsupported by your HW. No joke here, really.