v1.4.5rcX Bug Reports and Comments

[mike99]

192.168.1.0/255.255.255.0

If you mouse over the field it will show you that.

Didn't saw it.
Thanks

Default queue by port was removed because you can do that with a rule now.

That what I thinked first but was not able to find how except now with src-ip 0.0.0.0/0.0.0.0 as last rule.

[ap@dsp.lv]

Please add to watchdog notification about input voltage changes (fall)

As log cleared after reboot, please add option to send log to e-mail, before clear it.

If it is possible, please save statistic data over reboots (input voltage, current, efficiency)

Also can you watchdog add feature to monitor ports wattage (amps) to get know about PoE device when something wrong with power consumption.

Wonderful feature will be e-mail templating for notification with option to include time, switch and port current/wattage or other/all statistic data.
As there no easy to accumulate data from logs, even more that logs cleared after reboots, notification then can be customized as required with included all necessary data from that event moment. Also if supported to include tags as pictures - graphs with that time trends in past (10m, 1h, 24h).


Suggestion for additional hardware feature: add 1-wire port to monitor sensors on it. Of course with appropriate support in software - to monitor sensors with some action and notification.
1-wire can be from temperature, humidity, logical (DI, open/close), analog sensing (ADC) sensors till output for some action (digital output) for example to switch on cooling etc. even monitor battery charging/discharging.

https://www.maximintegrated.com/en/pl_l ... /filter/21
https://www.maximintegrated.com/en/prod ... 18S20.html
https://www.maximintegrated.com/en/prod ... S2450.html

[sirhc]

Please add to watchdog notification about input voltage changes (fall)

I assume you are talking about the SMART DC model of switches (WS-8-150-DC and WS-12-250-DC)?

We support full SNMP so you can monitor whatever you want and program your alerts to meet your needs.

If your talking about the WS-8-150-DC and WS-12-250-DC you can configure an alert to be sent when a low voltage is reached in the Power Tab.

As log cleared after reboot, please add option to send log to e-mail, before clear it.

Most all embeded devices lose their log when rebooted as the memory is cleared.

There is a button to email the log file on the Device/Log Tab.

You can also setup an external SysLog server on the Device/Configuration Tab so you can always have your logs.

If it is possible, please save statistic data over reboots (input voltage, current, efficiency)

This is why you run an SNMP and SysLog servers.

There is no way for us to keep statistics or Log data through a reboot.

Also can you watchdog add feature to monitor ports wattage (amps) to get know about PoE device when something wrong with power consumption.

We could do this but once again this is why we offer FULL SNMP support so you can do this in your SNMP server.

Make sure to running v1.4.5rc7 which has all the bugs reported fixed to date.

[sirhc]

Wonderful feature will be e-mail templating for notification with option to include time, switch and port current/wattage or other/all statistic data.
As there no easy to accumulate data from logs, even more that logs cleared after reboots, notification then can be customized as required with included all necessary data from that event moment. Also if supported to include tags as pictures - graphs with that time trends in past (10m, 1h, 24h).

These units have limited memory and CPU which is why we support SNMP and SysLog servers.
And everyone wants something different but everyone can get what they want by customizing their own SNMP server.

[ap@dsp.lv]

yes SMART DC model

> There is a button to email the log file on the Device/Log Tab.
> You can also setup an external SysLog server on the Device/Configuration Tab so you can always have your logs.
yes, right, it is solution with syslog.

> We could do this but once again this is why we offer FULL SNMP support so you can do this in your SNMP server.
yes, but it is require more complex system for small notification, but it is option.

> These units have limited memory and CPU which is why we support SNMP and SysLog servers.
template are not resource consumable feature, an it is give more flexibility to customize notification for fast "get known whats wrong" !
even template file can be without web front-end, so only staff with knowledge can modify (e-mail body) template files and upload to switch.
So, at least add switch timestamp to e-mail subj. or body.


Understood that switch have no internal clock.
Additional suggestion to buffer logs till update NTP time, shift log time to real (NTP) time.
Similar solution about year ago done Microtik (microtik.com) and now logs are more readable even after reboot .
I don`t know hove them do, save last time and then on boot restore it, or update NTP and shift log time - but logs no more have that records:
Jan 1 00:00:17

[sirhc]

I run a WISP myself and since all my infrastructure equipment is on invalid IPs I simply run my own internal NTP server which is very easy to setup then my internal non routable IPs have access to NTP server. My internal NTP server has a valid IP to get to world time servers and since it is inside my network my invalid IPs such as 172.17.10.2 can route to it to get time.

You can run an internal NTP server on Windows or Linux machine in your office
https://www.google.com/webhp?sourceid=c ... r+software

[wisphopefull]

+1 on running an internal NTP server, I have one running on a Raspberry Pi.

Are there still some memory leaks to be fixed?
I have 2 switches in production right now, WS-10-250AC (1.4.5rc7) and WS-8-150-AC (1.4.5rc4)
Memory usage is slowly rising all the time, but this was also happening on my WS-10-250AC when it was running 1.4.2.

All discovery services are disabled in the switches, the only thing enabled really is SNMP.

[sirhc]

+1 on running an internal NTP server, I have one running on a Raspberry Pi.

Are there still some memory leaks to be fixed?
I have 2 switches in production right now, WS-10-250AC (1.4.5rc7) and WS-8-150-AC (1.4.5rc4)
Memory usage is slowly rising all the time, but this was also happening on my WS-10-250AC when it was running 1.4.2.

All discovery services are disabled in the switches, the only thing enabled really is SNMP.

I am not noticing a memory leak.

Could you take a screen shot of the Device/Status page 1 day then at a later day/time another screen shot showing the increase in memory as well as maybe do a TOP screen shot at Linux level on both events then post the screen shots as that would help?

[wisphopefull]

Screenshots have been taken, I'll wait a while now and see.

In the meantime I have the graphs here from LibreNMS which is monitoring via SNMP.
The big drops in memory usage are due to me rebooting or upgrading the switches.

The first graph there is the WS-10-250-AC which was on 1.4.2 all the time until the drop which is when I upgraded to 1.4.5rc7.
And heres the config which is basically the same in both switches.

[sirhc]

LOL, not the Device/Configuration Tab, the Device/Status Tab that shows memory usage and version.