Netonix Forums

Hi.

Is answer for this situation? Memory usage is near 100 % (uptime 104 days). We have two same (WS-12-250-AC) switches on same location, one is high memory usage, second have only 52 MB. Configuration is basically the same.




Second switch have uptime only 4 days. Last reboot by watchdog ... It is maybe answer.

But we have third same switch with same firmware on other location, uptime 58 days and memory usage 53 MB.

Another example - firmware 1.4.9, uptime 193 days, memory usage 50 MB.

Please post up a screen grab of all the option on the Device/Status TAB

Possible causes:
Discovery TAB with large flat layer 2 segment? - Try disabling Discovery TAB on Device/Status TAB
Also Disable other Discover Protocols
SNMP polling to often, reduce SNMP polling times.
More than 1 Netonix Manager - Never use more than 1 Manager to monitor Netonix Switches.
Turn off SMTP for awhile see what that does.

Try disabling differnt services until you find what service is the issue.

Device/status TAB is first screenshot. I see no more values ...

Disabled all discovery (protocols and tab) I try on second switch. For now - memory is still 53MB (8 days uptime). It is not large layer2 segment. Every connected rocket have own VLAN. Only about 35 devices (visible in discovery tab) on one switch.

SNMP pooling is not possible reduce. A few get and walk twice in hour (it is my own management system). No Netonix manager, no zabix, and so on ...

At this point it seems to be a problem with device discovery.

Device/status TAB is first screenshot. I see no more values ...

Sorry I meant to request the Device/Configurations TAB

At this point it seems to be a problem with device discovery.

Please elaborate, are you saying it is the Device Discovery TAB being enabled or a discovery protocol, and if so which one?

My english not good ...

First switch - screenshots in this thread, all discovery protocols enabled. Discovery tab enabled too.
Second switch - a week ago rebooted by watchdog. After that all discovery protocols disabled, discovery tab disabled too. Memory seems ok.

Please Enable the Discovery Protocols 1 at a time and see.

I think it is the Discovery TAB not the protocols but we will see.

Then go from there.

First switch:

turn off discovery tab = memory usage no change
turn off ubnt discovery = no change
turn off lldp = no change
turn off cisco = no change

turn off Loop Protection = memory usage to 42 MB!

After that I turned on everything back - memory usage still 42.

On second switch I turn on Ubnt discovery and discovery tab. No change in memory usage.

There is a short time for verdict, of course.

----
IMHO it is problem version 1.5.0. For example:
I have switch with version 1.4.9. All discovery options on (24 devices). Uptime 241 days. Memory usage 54MB.
Other piece - 119 days uptime, 25 devices in discovery. 49MB memory usage.

Other ws-8-150-AC with version 1.5.0 - uptime 49 days, 34 devices in discovery. 94MB memory usage.

Other ws-8-150-AC with 1.4.9 - uptime 70 days, 12 devices in discovery. 47MB memory usage.
Other ws-8-150-AC with 1.4.6 - uptime 679 days, 24 devices in discovery. 54MB memory usage.
Other ws-6-mini with 1.4.8 - uptime 68 days, 28 devices in discovery. 47MB memory usage.

Loopback detecion is on in every my switches.

Try v1.5.1rc8 let me know

Unfortunately, company policy does not allow us to deploy beta / rc in production. Do not even make other changes on the network during Christmas holidays.

Switch "Other ws-8-150-AC with version 1.5.0" rebooted by watchdog today. I don't know memory usage at this moment :-( But at first I try disable loopback-detecion. In changelog 1.5.0 you wrote "- Improved ability to deal with loops while loop protection".

Will see after one or two month.

Ludvik wrote:Unfortunately, company policy does not allow us to deploy beta / rc in production. Do not even make other changes on the network during Christmas holidays.

Switch "Other ws-8-150-AC with version 1.5.0" rebooted by watchdog today. I don't know memory usage at this moment :-( But at first I try disable loopback-detecion. In changelog 1.5.0 you wrote "- Improved ability to deal with loops while loop protection".

Will see after one or two month.

As explained many times:
Say we release v1.5.1 TOMORROW
Then on Monday someone reports a bug.
If we verify the bug we then release v1.5.2rc1 which is a BUG FIX

If you can not use RC versions and your issue is a bug in v1.5.0 then I guess you have to wait and suffer until we finalize v1.5.1

The only reason we finalize something is because we feel like saying OK time to finalize so we can provide to SMT shop to pre install on production boards.

I argued with programmers in the beginning and wanted to call it bf instead of rc but I lost the argument.