GS116Ev2 / GS108Ev3 losing management IP if dhcp server is not reachable after 2 hour renewal period
There seems to be a bug within the dhcp client of GS116Ev2 (Firmware Version 2.6.0.43) and GS108Ev3 (Firmware Version 2.06.14). If the switch is configured to get its management IP address from a local dhcp-server and this server is (for some reason) not available after the renew cycle of 2 hours (probably hard coded) , the dhcp client in the switch stops trying to get a new lease leading the switch in an unmanagable state. I observed this behavior with two GS116Ev2 and two GS108Ev3 several times when our dhcp server was down for maintenance during the renewal attempt .
Obviously the build in dhcp client has got a fixed interval of 2 hours trying to renew the IP address for the management interface, because the given default-lease-time of the dhcp-server is simply ignored (I tried several settings). If the dhcp-server is not reachable the dhcp client in the switch seems to die instead of trying to discover a new lease. Since the Switch Management Mode is set to Web browser only since second last firmware version there is no chance to get the management IP back until the power plug is pulled once.
Intended behaviour:
If the build in dhcp client does not get a lease from the dhcp-sever the client should keep on trying until it succeeds.
How to reproduce:
Configure one of the named switches as dhcp client. Shutdown the dhcp-server for more than 2 hours. After 2hours + x the management IP of the switch is no longer reachable. Only pulling the power plug brings the interface back.
Can anybody confirm this behaviour and drop a note to the developers?
Model: GS116Ev2|ProSafe Plus 16 ports gigabits switch
If the switch is configured to get its management IP address from a local dhcp-server and this server is (for some reason) not available after the renew cycle of 2 hours (probably hard coded) , the dhcp client in the switch stops trying to get a new lease leading the switch in an unmanagable state.
The dhcpd involved is more than basic, I think you are right ref. the two hours. It does not stop - much more the switch will fall-back to the default IP address. Try the Netgear SWitch Discovery Utility - this might confirm the default IP...
Obviously the build in dhcp client has got a fixed interval of 2 hours trying to renew the IP address for the management interface, because the given default-lease-time of the dhcp-server is simply ignored (I tried several settings).
Since the Switch Management Mode is set to Web browser only since second last firmware version there is no chance to get the management IP back until the power plug is pulled once.
You are the network admin. Netgear has (as per the release notes)"only" enforced a new default. You are free to enable the NSDT controls again.
If the build in dhcp client does not get a lease from the dhcp-sever the client should keep on trying until it succeeds.
Here again, the default behaviour of the tiny controller does jump in: No DHCP -> fall-back to the the default LAN IP config. For the initial set-up and the first attempt after power-up this is correct. For renewal, the simple default is wrong IMHO.
Configure one of the named switches as dhcp client. Shutdown the dhcp-server for more than 2 hours. After 2hours + x the management IP of the switch is no longer reachable. Only pulling the power plug brings the interface back.
Correct. Under normal operation conditions, this should never happen...
We will fix the DHCP client issue in new models, but not in old models.
In the light of the limited lifetime warranty, and because this is a major flaw, old would be devices which past the EoL announcement more than five years ago. Anthing else must be considered new. @YeZ please.
Sorry. Please use plus utility to renew your switch IP.
Do I understand your suggestion right, that we should enable NDSP although in Firmware Version 2.6.0.43 the plus utility was disabled by default (due to security reasons)? This advice can not be serious, right?!?
We will fix the DHCP client issue in new models, but not in old models.
Old models? GS116Ev2 (and GS108Ev3 - same problem) are current models still available and sold on the market?!?
I think Netgear needs some kind of change in strategy! In view of the current problems, I can no longer recommend Netgear "business hardware". For DHCP issue I would have to buy new hardware model which is currently unavailable (mine is not even 2 years old).
The latest Firmware update 2.6.0.48 uses a minimum MTU of 1500 (see German Forum) so that maintainance through Webinterface via VPN is not possible anymore. Some devices are not even reachable with pro utility after firmware update. All this without notifying customers in advance, and accepting to break well established setups? Furthermore the one and only response to customers who investigated time in tracking down problems is either to say we'll not fix - please buy next generation model- or thanks for tracking down the issue "corresponding information will be published on our pages in the next few days" but for security reasons you "should now set the MTU value to at least 1500 so that HTTP access can work properly"?
To sum up: Firmware of GS116Ev2 and GS108Ev3 is totally broken by desing. Instead of providing an adequate fix a crappy workaround is provided. This is unworthy of being called business hardware.
Maybe its time to distibute this approach to a wider range using social networks and professional journals / websites.
The latest Firmware update 2.6.0.48 uses a minimum MTU of 1500 (see German Forum) so that maintainance through Webinterface via VPN is not possible anymore. ... Furthermore the one and only response to customers who investigated time in tracking down problems is either to say we'll not fix - please buy next generation model- or thanks for tracking down the issue "corresponding information will be published on our pages in the next few days" but for security reasons you "should now set the MTU value to at least 1500 so that HTTP access can work properly"?
It's something I could understood decade(s) ago where NSTP had to pack "everything" into one frame (this problem was and is real!) because of packet fragmentation was (and is!) not possible due to the very simple embedded controller architecture and it's hmm substandad (fair enough?) IP stack, and DHCP implementation, ...
@YeZ this is a completely unacceptable change, even in the eyes of "urgent" security fixes. Minimum MTU workable should be some 1400 to cover every possible use case.
Betreff: GS116Ev2 / GS108Ev3 losing management IP if dhcp server is not reachable after 2 hour renew
Hello all together,
does the problem still exist? It looks like my new GS308E has the same behavior. In this case the swicht is not usable for me, because behind the swicht is an important part of my smart home. And when this siutaion happens, my smart home works not correct anymore. An at the moment it looks like i have the same problem.
Kind reagards
Dennis
Model: GS308E|8 Port Gigabit Ethernet Smart Managed Plus Switch
does the problem still exist? It looks like my new GS308E has the same behavior.
What problem, what behavior?
This thread does talk about the way the DHCP client (for the local switch management) and the minimum MTU required to access the management Web UI. Undoubted, the DHCP implementation is still the same; no idea if the minimum MTU enforcement prohibiting the switch Web UI access on smaller MTU conditions (like through a VPN) is in place on your switch model and the undefined firmware version.
In this case the swicht is not usable for me, because behind the swicht is an important part of my smart home. And when this siutaion happens, my smart home works not correct anymore. An at the moment it looks like i have the same problem.
There is no problem in this thread that is happen suddenly and making the switch no longer workable in general.
does the problem still exist? It looks like my new GS308E has the same behavior.
What problem, what behavior?
The behavior of th DHCP. My Problem is, that when the DHCP-Server is not reachable, the GS308E lost the static IP. Thus, the switch and all devices behind the switch are no longer accessible. That is a big Problem in my private network. Most of the time this happens when I am not at home and therefore I can't pull a plug. During this time, my smart home does not run correctly.
Or am I misunderstanding this thread?
Greetings Dennis
PS: My swicht is the GS308E with the actuall version V1.00.08GR.
Model: GS308E|8 Port Gigabit Ethernet Smart Managed Plus Switch
The behavior of th DHCP. My Problem is, that when the DHCP-Server is not reachable, the GS308E lost the static IP. Thus, the switch and all devices behind the switch are no longer accessible.
... Or am I misunderstanding this thread?
Yes, I think you do. This does only affect the management of the switch. The other L2 traffic, including any communication to your IoT isn't affected by this issue at all-
