TL-SG108PE Factory resets every power cycle & weird switching issues
This thread has been locked for further replies. You can start a new thread to share your ideas or ask questions.
Hi all,
I purchased the TL-SG108PE smart switch a couple of months ago and running into two issues:
(1) Every time I power cycle the switch, it factory resets, reverting to the factory default IP 192.168.0.1. This is highly inconvenient; is it standard behaviour/is there a way to set a persistent configuration or does this sound like a fault?
(2) I had a strange issue today. My network architecture looks like this: WAN > pfSense Router > TL-SG108PE Switch > NAS/WAP (same subnet). A couple of hours ago I found I couldn't connect to any devices on the network, with the exception of the router and the bare metal FreeNAS instance on the NAS. I was *not* able to connect to any of the VMs hosted within the FreeNAS jail, OR any other devices connected via the WAP (4-5 laptops, phones, tablets etc.). Pinging these devices resulted in a "Time to live exceeded" error. Performing a traceroute on the same hosts showed that it was connecting to external servers, and made it about 6-7 hops before converging on two external servers and oscillating between them. I'm assuming this was occuring because the hosts couldn't be found through the switch, so the traffic was routed to the router to find them. Power cycling the switch rectified this issue. My question is what could cause this behaviour and how can I ensure it doesn't happen in the future? The only settings I had changed were to change the switch from static IP reservation to DHCP to be allocated an IP by the router.
Any help is appreciated.
Cheers.
Did you save the config after applying changes? If not, factory defaults will be restored at next boot. Note that older firmware had the problem to only save the config once and ignoring all other requests to save again after more changes had been applied. If you have this buggy firmware (can be seen on the missing "Saved config to flash" success message after repeatedly saving the config), then either reboot the switch each time after a config save to ensure next save will work or upgrade to the latest firmware.
As for the second problem could have to do with WoL. While my MacBook's NIC wakes up when it receives a ping packet, other systems might need a magic packet.
Re: DHCP for stationary devices: don't do it. DHCP is fine for mobile devices, but not for stationary devices, especially if you need to access those devices. At lest, assign an IP using a DHCP MAC-to-IP mapping if you insist to use DHCP for each and every device in your network (except for the router I hope 
).
R1D2 wrote
At lest, assign an IP using a DHCP MAC-to-IP mapping if you insist to use DHCP for each and every device in your network (except for the router I hope
Yeah, this is what I'm doing. My thought is that it's less confusing to be consistent about how IPs are allocated - better to have the router dictate everything rather than allow individual devices to specify their own ip. Means I can make adjustments all from one place as well.
As for saving the config, I'm 99% sure I did but the last time I power cycled and had to set everything up again was over a month ago so I don't have a clear memory. I'll make sure I actively save it this time.
RE: WoL, I don't think this is the issue. It doesn't seem likely that this would affect the majority of my devices all at once.
Mitya: If it isn't a L2 issue, why would it have been resolved by power cycling my L2 switch? I wasn't experiencing problems pinging the router or accessing the internet, only other devices on my LAN. Doesn't this indicate an issue with the switch? Interestingly, whilst the NAS was accessible, the IPMI for the NAS was not. Both are connected to the same switch, so I'm a bit confused why one would be accessible but not the other