Final 6.2.1 update causing LAN instability/sluggishness

StephenB wrote:

dsnpevl wrote: StephenB wrote: I agree that a list of routers could be helpful for them. I don't think this is likely to have much to do with the chipset. More likely it relates to the IP software stack in the router. Don't IP protocol stacks run inside chipsets? Software can and is ported to more than one platform. Platforms can often run more than one software stack. So a software bug can affect multiple routers using different hardware chipset, and there might be other routers using the same chipset that have no problem at all.

OK. It's even more complex than I thought. And comparing chipsets in routers would not bring conclusive evidence of the problem, I guess.
So if I understand this correctly, it could be the connman 1.27 in RN in combination with a certain version of a certain IP software stack inside the router that don't particularly work well together?
That will be pretty hard to trace back. Guess we'll stick to the router list then :)

I want to add what I have discovered myself, in case it is helpful here.

I started having issues this morning after I updated our ReadyNAS 314, but I didn't connect the issues I have having with the update until much later. First our internet went out, which I thought was our ISP, but I was able to connect with a single system bypassing our router. I plugged the router back in and ran a packet capture an noticed our entire upload bandwidth was saturated with OpenVPN packets to our offsite salesman. I disabled our OpenVPN server and the internet came back to life.

Around the same time I was notified of a printer issue. One of our wireless printers wouldn't talk to the network, and I couldn't log in to the WAP that it should connect to. So I mirrored this port to my computer and ran another packet capture. I noticed the network was being flooded with DHCPv6 requests, thousands every second. I don't run an IPv6 network, and don't even have server support for it at this time. Some more digging tracked it to an IPv4 address in my DHCP pool range. Even more digging tracked it back to my ReadyNAS MAC, which also had a static IPv4 address. I am only using one network port on this, so I don't know why it would have 2 IP addresses, or what it is trying to do with IPv6. I had to unplug it to bring my network back, but now I don't have my ReadyNAS, my main storage archive.

The only app I have installed is Anti-Virus Plus. I don't run any media servers. If I can provide any more info, let me know and I will. I hope to get this tracked down soon.

This may be a bit off topic, but I believe it is related to the 6.2.1 update. Hope it helps in finding a solution.

After my update I did not seem to have a problem. My problem occurred after installing then uninstalling the application 'Contacts and Calendars Server'. I can now no longer log into the admin web page at all. When I try to get to it via a browser, it says it does not connect. There is no login window that pops up, nothing. I have no access to controlling the RN104 any more. I can see my files on the various shares but but I cannot access the admin interface at all.

Again, I apologize if this is off topic and should be somewhere else. I'm a newb to forum posting. Hope this helps in the list of symptoms and a finding a solution. :oops:

EDIT: Optimus prime moved this the new thread http://www.readynas.com/forum/viewtopic.php?f=65&t=79158

EDIT: Thanks Optimus Prime. -- Jack_A_Lope :)

@Jack_A_lope, could be related. I'm just curious, did you do a reboot after the uninstalling of the application? For me the upgrade works fine at first install, but after reboot, the problems kicked in. I'm wondering if this also applies for you?

My meaning was not to infer issue is related to one brand of router.

For now this is information we have for those who view this thread.
Symtoms:
After upgrade to 6.2.1 and reboot
Network instability seen and some services cannot be accessed on the NAS
Can access ReadyNAS GUI but the network ports do not show on Network tab.
Can see flooding on the network. (synchronised fast blinking LED on router / switch)
May cause issue with home network not being accessible. Unpug Network cable from ReadyNAS solves home network issue

Information known so far
Issue is related to IPv6 and latest version of connman (connection manager 1.27) used by 6.x to manage all network interfaces.
This causes flooding of IPv6 information request packets.
We can reproduce with some routers and not others which are running DHCPv6.

Solutions
If not using IPv6,

unplug ReadyNAS
disable IPv6 DHCP on the router
reconnect ReadyNAS

If need IPv6,

unplug ReadyNAS,
disable IPv6 DHCP on the router reconnect ReadyNAS,
Down grade connman to 1.26 using SSH access.


apt-cache showpkg connman


check 1.26 is listed then run


sudo apt-get install connman=1.26-netgear1


enable IPv6 on router
Reboot ReadyNAS

These steps are basic listed by me from basic testing done in house there maybe other ways to achieve this.
I see other issues reported in this thread but if symptoms not like above then please start a new thread to get help from the community.

Netgear is working towards a release to solve this issue

6.2.2-RC1 is now available and has connman downgraded back to 1.26: viewtopic.php?f=154&t=78866

If some of you are able to try 6.2.2-RC1 that would be appreciated.

If you are currently experiencing the issue, follow these steps:

1. Unplug ReadyNAS
2. Disable IPv6 DHCP on the router
3. Reconnect ReadyNAS
4. Install 6.2.2-RC1
5. Enable IPv6 on router
6. Reboot the ReadyNAS once more.

mdgm wrote:
If some of you are able to try 6.2.2-RC1 that would be appreciated.

Followed suggested procedure. After install did a shutdown / powerdown and powerup. Next I tried restart from Web Admin page.

None of the symptoms I saw with 6.2.1 final have emerged so far. So I think this 6.2.2-RC1 is working properly and fixes the "LAN instability/sluggishness" mentioned in this thread's title. Thanks to everyone involved for providing the fix at such short notice!

One point of attention, that I think I already mentioned somewhere in this thread, when I did a re-install of the OS 6.2.1 via the Reset+PwrOn/boot menu. In the Web Admin, the timer bar for re-establishing the connection runs out before the update is fully installed. This also happened when following the 6.2.2-RC1 installation instructions above.



I get a "Connection Lost" message as depicted below, when on the front display of the RN, the message "booting" is still visible. When the front display show the OS version 6.2.2, a simple reload of the page will bring back the Web Admin page. So the connection lost message pops up too soon. That might confuse users and make them want to do a manual reboot, while actually the installation is succesful. It just toke a bit longer then expected.



UPDATE 2014-12-24: Saw the update 6.2.2 (final?) and installed it. No network issues at first glance. Can still connect and no escessive blinking on the switch.
But did get the Connection Lost error message again as described above. For me the Device Offline timeout would need to be aprox. 30 seconds longer, to avoid this issue.

I manually downgraded to connman 1.26 last night pre mdgm RC1 post via SSH, I did not reboot following the re-install. Post the downgrade the router remained accessible as did the wider internet.

This morning the NAS (RN102) auto powered down, post waking the unit up the same issue was evident as before. The router was inaccessible as was the internet. Upon switching off the NAS, the router became accessible without restart as did the wider internet.

I have performed the upgrade to RC1 however this has not remedied the problem. The router is still inaccessible as is the wider internet. The NAS is accessible though.

My main router BT HomeHub 4 does not have the option readily available to stop or turn off IPv6, I have moved to the backup router TP-LINK TD-8817 which does have this option however the router and the internet does still become unavailable almost as soon as the NAS is switched on, the NAS itself is still accessible.

I do not have a great deal of expertise in terms of packet capture of analysis so I have the printouts below in the hope they shed some light -

ifconfig printout - (The NAS had just finished booting when the below was run, I have no applications running bar Anti-Virus Plus)

eth0      Link encap:Ethernet  HWaddr xx:xx:xx:xx:xx:xx
          inet addr:192.168.1.XXX  Bcast:192.168.1.255  Mask:255.255.255.0
          inet6 addr: XXXX::2ac6:8eff:fe34:681/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST DYNAMIC  MTU:1500  Metric:1
          RX packets:2054 errors:0 dropped:0 overruns:0 frame:0
          TX packets:9596834 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:512
          RX bytes:403184 (393.7 KiB)  TX bytes:1589665015 (1.4 GiB) 
          Interrupt:10

Should the unit have an inet6 address given that the router shouldn't be giving them out, or is this an auto configured IP?

The TX bytes figure seems huge (my internet is barely 4mbp/s, and I had nothing over the LAN bar the SSH connection)

I also ran tcpdump, I didn't save the capture file unfortunately but can run again if it is any use.

tcpdump -i eth0 --vv

2493 packets captured
3155450 packets received by filter
3152926 packets dropped by kernel

I really have no idea about the packets received or dropped, I am off to Google that.


I hope some of the stats help I am just about to email my logs over and am more than happy to run anything else that would help to assist and solve this problem.

jimbo123 your disk's SMART stats look bad though they haven't changed in some time. Can you reproduce this issue if you use a disk from the compatibility list?: http://kb.netgear.com/app/answers/detail/a_id/20641

For the IPv6 address showing

inet6 addr: XXXX::2ac6:8eff:fe34:681/64 Scope:Link

That is known as link local and exists the moment IPv6 is enabled

The packet trace if can take a small window (not huge trace) and send in us same process as send logs would help see what is generating the packets and if really is same issue.

The TX / RX size seems large if you were only rebooted short time.
For example unplug and replug cable does not rest the counter.
Going to auto shutdown and waking up may not reset and you would need a full reboot cycle I think, but must confirm.

Could you try and OS reinstall when on 6.2.2 RC1, this will reset network settings and admin password.