v1.3.8 FINAL bug reports and comments

[designweb]

thanks for adding the timer, it all gets better and better, upgraded 8 switches so far 1.3.8 and all have log showing as below but all have perfect time in the top of the gui and cli

Jan 1 00:00:49 system: starting ntpclient
Jan 1 00:01:00 system: starting ntpclient
Jan 1 00:01:11 system: starting ntpclient
Jan 1 00:01:22 system: starting ntpclient
Jan 1 00:01:33 system: starting ntpclient
Jan 1 00:01:45 system: starting ntpclient
Jan 1 00:01:56 system: starting ntpclient
Jan 1 00:02:07 system: starting ntpclient
Jan 1 00:02:18 system: starting ntpclient
Jan 1 00:02:29 system: starting ntpclient
Jan 1 00:02:40 system: starting ntpclient
Jan 1 00:02:51 system: starting ntpclient
Jan 1 00:03:02 system: starting ntpclient
Jan 1 00:03:13 system: starting ntpclient
Jan 1 00:03:24 system: starting ntpclient
Jan 1 00:03:35 system: starting ntpclient
Jan 1 00:03:46 system: starting ntpclient
Jan 1 00:03:58 system: starting ntpclient
Jan 1 00:04:09 system: starting ntpclient

[sirhc]

thanks for adding the timer, it all gets better and better, upgraded 8 switches so far 1.3.8 and all have log showing as below but all have perfect time in the top of the gui and cli

Jan 1 00:00:49 system: starting ntpclient

[ 17 ATTEMPTS TRIMMED ]

Jan 1 00:04:09 system: starting ntpclient

This is obviously showing the NTP client trying to acquire the time from the NTP server. Not sure why it took 3 minutes?

This can be caused by your network topography if you are using higher level routing protocols such as OSPF until the adjacency come up and actually allow the switch to access the NTP server?

At my WISP we run a local NTP server but still if the tower uses OSPF it sometimes takes a minute or 2 for the OSPF Adjacency to come up and then for the network to converge its routes to allow the NTP server to be accessed.

[designweb]

We do run OSPF, however on the next three switches I have just done have nothing in the logs and all working well, I had not noticed this in any previous upgrade so yes probably adjacency took a while. Next step Ethernet Automatic Protection Switching (EAPS) and Extreme Discovery Protocol and these will be the perfect switch to talk to our network core. A Very Happy User

[alohabro]

Upgrade the DC switch at my house with no issue: WS-8-250-DC
I'll role this out to my active switches next week.

[reighbread]

Just upgraded my WS6 to 1.3.8. Everything looks great so far, if I don't post back then you guys are doing a great job!
The only thing I noticed was that the Rocket plugged into port 6 did a power cycle on upgrade, but ports 3, 4, and 5 did not. No big deal, but just interesting.

What does the switch log say about port 6?

Nov 21 09:21:01 syslogd started: BusyBox v1.19.4
Nov 21 09:21:02 dropbear[1318]: Running in background
Nov 21 09:21:04 kernel: link state changed to 'down' on port 6
Nov 21 09:21:04 sysinit: 981
Nov 21 09:21:07 switch[1343]: Detected warm boot
Nov 21 09:21:10 kernel: link state changed to 'up' on port 6

[bmv]

Upgraded 3 switches from various prior versions, 1.3.2 and 1.2.x and all good so far.
Got another dozen or so to upgrade when time allows.

6-MINI all okay.
8-DC all okay
24-A all okay.

Just works! :)
No issues to report after 24 hours.

Thanks for the good work.

[LRL]

I've upgraded 3 ws-12-250-DC switches, and a ws-8-250-DC and all have been fine!

Chris, it even fixed the switch i was having all those issues with voltage and watt readings!

Here's a suggestion, I personally graph my DC voltage using Mikrotiks, but it would be nice if the switch kept a graph history like it does with Total Power Usage

[bmv]

I've upgraded 3 ws-12-250-DC switches, and a ws-8-250-DC and all have been fine!

Chris, it even fixed the switch i was having all those issues with voltage and watt readings!

Here's a suggestion, I personally graph my DC voltage using Mikrotiks, but it would be nice if the switch kept a graph history like it does with Total Power Usage

You're best to do this via SNMP.

We use PRTG and when you import the Netonix MIB file, you can specify all the voltages and then graph these as different sensors with full historical reporting and alerting.

We have this working on some of our critical switches.

[LRL]

Don't get me wrong I graph our voltages from SNMP, but just thought it would be a nice feature if the switch did it as a built in.

I have found the voltage reading from the switches are more unpredictable due to the resistance that occurs because of the load of the radios. I base alarms off of voltage, but I need a true reading from the battery without resistance lowering it. This is the only reason we use another device. We do use the switch reading to verify should we have an alarm.

Bare in mind that I run all of my sites with 12volt banks which leads to lower voltages and higher amp draws and as a result more resistance and more voltage drop along the path to the switch.

[Banana Jack]

Upgraded five WS-6-MINI switches, all looking sweet. Keep up the great work! Thanks.