Netonix Forums

mayheart wrote:Expanding on the SFP issue.

I ran into this today. copper SFP on a 12 port AC switch.

I made a vlan change and the port went down. This was in the log until I manually shut the port down on the other side and brought it back up.

FWIW at this point - We just ran into this issue as well for the first time now on a WS-8-150-DC running 1.5.14. Unable to resolve this remotely, we have replaced the switch to investigate the problem one in our lab.

Our remote sys logs:

Code: Select all

Jun 22 09:28:09 10.51.242.194 :: Jun 22 09:28:09 10.51.242.194 UI: Configuration changed by root (::ffff:10.1.96.18)
 Jun 22 09:28:09 10.51.242.194 :: Jun 22 09:28:09 10.51.242.194 UI: VLAN 164 PortSettings: changed from 'EEEEEEET' to 'TEEEEEET'
 Jun 22 09:28:09 10.51.242.194 :: Jun 22 09:28:09 10.51.242.194 UI: VLAN 184 PortSettings: changed from 'EEEEEEET' to 'TEEEEEET'
 Jun 22 09:28:16 10.51.242.194 :: Jun 22 09:28:16 10.51.242.194 STP: msti 0 set port 6 to learning
 Jun 22 09:28:17 10.51.242.194 :: Jun 22 09:28:17 10.51.242.194 STP: msti 0 set port 6 to forwarding
 Jun 22 09:28:17 10.51.242.194 :: Jun 22 09:28:17 10.51.242.194 STP: msti 0 set port 5 to learning
 Jun 22 09:28:17 10.51.242.194 :: Jun 22 09:28:17 10.51.242.194 STP: msti 0 set port 5 to forwarding
 Jun 22 09:28:17 10.51.242.194 :: Jun 22 09:28:17 10.51.242.194 STP: msti 0 set port 4 to learning
 Jun 22 09:28:17 10.51.242.194 :: Jun 22 09:28:17 10.51.242.194 STP: msti 0 set port 4 to forwarding
 Jun 22 09:28:17 10.51.242.194 :: Jun 22 09:28:17 10.51.242.194 STP: msti 0 set port 3 to learning
 Jun 22 09:28:17 10.51.242.194 :: Jun 22 09:28:17 10.51.242.194 STP: msti 0 set port 3 to forwarding
 Jun 22 09:28:28 10.51.242.194 :: Jun 22 09:28:28 10.51.242.194 SFP: I2C communication down to SFP 1, attempting to establish communications
(END)
 

hello any update that corrects the crazy use of memory by monitoring? we have more than 90 netonix Ws with problems and this is simply worrying that we do not have an update to correct the errors, it is unheard of. snmp can not be a solution to disable it as well as schedule reboots. there must be a technically court and common sense solution.

jpmoller wrote:For any others with the memory leak issue with the snmp server, we've created a cronjob that evaluates the memory at 3am. If available memory is low, it restarts the server. If that doesn't fix the memory issue, the host is rebooted.

If you're interested, flick me a message. Happy to share it with the community.


Fun fact, even though cron is in the list of services that are meant to be started at boot, it does in fact not start up. I should be surprised, but alas...

Would be interested in seeing what you've come up with. Have also found that the vtss_appl process is bad for memory.

I have a WS-250-DC running 1.5.14 that apparently has a VLAN issue as well. Have a Ubiquiti Rocket2/Prism AP (firmware 8.7.11) VLANed to a local Mikrotik 3011 that randomly looses its VLAN connection. Still up on 24v POE power, still has 1G Ethernet connection, but no longer pingable and not passing traffic. Have replaced the AP three times, the network cable three times, and the switch twice (both running 1.5.14. POE bouncing the port does not correct the issue nor does rebooting the switch. Moving the AP to another port sometimes does, but it will occur again on the new port. When it does occur, it will eventually for no particular reason begin passing traffic again sometimes hours later.

This same AP switch combo ran for 3 years with no issues. Is their an earlier version of firmware where this is known to not be a problem? I went from 1.5.14 from 1.5.8 on this switch, never ran the intermediate versions.

Thanks, Doug

Just popping a note here that v1.5.15rc1 dropped 5 days ago which says to address the SNMP memory leak, as well as watchdog issue. We didn't get a notification, so I thought I might post here in case others also didn't see
https://forum.netonix.com/viewtopic.php?f=17&t=7869

We are running this on a lab switch with known memory leak for the last few hours, and it looks promising already. Will re-evaluate after a week.

v1.5.15rc2 to drop soon.

rc1 fixed most memory leak issue (much better) but an additional leak in log was found thus rc2