I upgraded to release 6.9.0 and started getting the following message from my ReadyNAS 516.
"Bond interface bond0 has slave interface eth0 offline."
I went to look at the NAS (RN51600-100NAS) and sure enough the LINK lights are going off and on (on both the NAS & switch). The message is legit that the link is going up and down. I tried the following with no sucess:
* Swap to a different port on switch (Dell 2448)
* Change the actual cable
* Switch the cables eth0 and eth1
* Reboot
Go back and look at my device logs off my network management tools (Ubiquiti and FingBox) and this has been happening for a long time so 6.9.0 just exposed the isssue.
There have been many reports of problems with the network interface on OS6.9.0, and I believe yours is related to that. It seems like a diver problem which can cause the ethernet subsystem to reset periodically. I would hold off on condemning the hardware just yet,
There have been many reports of problems with the network interface on OS6.9.0, and I believe yours is related to that. It seems like a diver problem which can cause the ethernet subsystem to reset periodically. I would hold off on condemning the hardware just yet,
So I think you nailed it, this morning both my ReadyNAS 212 and 516 were messed up. The 212 now reports having a bad port eth0 and now my 516 doing a resync. This release 6.9.0 was not ready for primetime. I think I may need to downgrade to protect myself....I guess I have learned my lesson. Don't trust Netgear to do QA.
Thanks Sandshark!
Model: RN21200|ReadyNAS 212 Series 2-Bay (Diskless)
Yea, I have trouble with LACP as well, running 6.9.0. There are alot of posts regarding it, and I have a ticket open with support regarding this as well.
In another thread @mdgm-ntgr has warned that he believes there is something in 6.9.0 that should not be downgraded. It's not just a case of them always including that statement in the release notes. So be prepared for the possibility of needing to factory default if you do downgrade.
I have been increasingly disappointed in what appears to be poor software testing on the part of Netgear. The appearance is that they rely too heavily on beta testing instead of having a formal test process in house or that they have poorly documented requirements. Specifically, they have poor retrograde testing (showing a new feature didn't harm an old one). Testing is only as good as the requirements, because you generally only test the requirements. If Netgear management is driving this as a cost savings or to push out shiny new features before they are ready, they are ruining their reputation in the process. If engineering is driving it, then just shame on them.
I recognize there are a lot of configurations that have to be tested. Intel and ARM. Different RAID configurations. Active directory or not. With or without expansion chassis. With or without port aggregation. And so much more the permutations become too vast to test them all. But many of the bugs I have seen reported do not appear to be in wildly exotic use cases.
I am never a first adopter of a 6.x.0 release except on my "sandbox" Pro2 system because of Netgear's track record for quick 6.x.1 (and in some cases .2, .3, etc.) to fix bugs.
