v1.5.25 Bug Reports and Comments

[mayheart]

Weird SNMP bug I encountered.

SNMPv2-MIB::sysUpTime.0 is always returning 1843777931 after a fresh upgrade to 1.5.25.

If you reboot the switch one more time, it returns expected values.

Is this a one off or every switch you upgrade and from what version was the upgrade.

I'm seeing it on every switch.

One was 1.5.24 to 1.5.25

The other 3 were 1.5.15rc3 to 1.5.25

[oeyre]

I'm seeing it on every switch.

One was 1.5.24 to 1.5.25

The other 3 were 1.5.15rc3 to 1.5.25

Do you have to reboot or can you disable and enable SNMP in the device configuration?

We noticed an issue with SNMP not updating traffic counters for approx 15% of our install base when upgrading from 1.5.16 to 1.5.22 which was resolved by rebooting or toggling SNMP.

[sirhc]

Once upgraded to 1.5.25, a reboot "may" be required for a "few" especially when upgrading from v1.5.17rc2 or older.

But once upgraded to v1.5.25 things should be fine.

This was a major re-write 60K lines of coding to remove PHP.

There may be an issue with the manager but waiting on some more feedback back. If there is an issue with the new version(s) it will require an upgrade to the manager not the firmware as the firmware was changed for security.

For now as we think we have solved all reported issues although a small number may need to reboot after upgrade and possibly a very small number to power cycle because of the massive software architecture change

For now we have the coders back on the WS3 product as we are now 4 months behind on that due to the hack and while we did the rewrite we fixed and improved a lot of long standing issues like some SFPs not coming back up after reboot. But for the record Cisco brand always did but they follow the standards sadly they cost a lot more.

But after we find out if the problem with the manager is confirmed which is possible due to session fix done in v1.5.25 to fix that false drop down red box false reporting a reboot had occurred. Again caused by making things more secure.

[yahel]

@sirhc and Netonix team THANKS!

Sorry for not reporting earlier, it's been a crazy day...
Vivek upgraded a single switch that was having problems on 1.5.24 (OSPF via SFP not working) - and he reports that the problem is SOLVED!

I have not yet checked anything myself - but it sounds good!
We will upgrade more and test more tonight.

I do get an error for that switch in Netonix-manager: "Error updating device"...

More later...

Thanks again !!

[yahel]

Update - ignore my previous report about the manager (I had a stale HTTPS session).
The only problem with Netonix-manager is that when using the manager to open the web-gui of a switch it gets "403 Forbidden".
That's been the case since 1.5.21 --- not a big deal at all... Can certainly wait...

More later...

[sirhc]

Update - ignore my previous report about the manager (I had a stale HTTPS session).
The only problem with Netonix-manager is that when using the manager to open the web-gui of a switch it gets "403 Forbidden".
That's been the case since 1.5.21 --- not a big deal at all... Can certainly wait...

More later...

When we get more feed back on manager we will release a new version to fix issues.

[sirhc]

I'm seeing it on every switch.

One was 1.5.24 to 1.5.25

The other 3 were 1.5.15rc3 to 1.5.25

Do you have to reboot or can you disable and enable SNMP in the device configuration?

We noticed an issue with SNMP not updating traffic counters for approx 15% of our install base when upgrading from 1.5.16 to 1.5.22 which was resolved by rebooting or toggling SNMP.

So to confirm, some (85%) switches when upgraded SNMP worked fine and some (15%) a reboot or toggle fixed it?? -- KINDA WEIRD???

If above is correct it would be helpful to "try" and figure out why, maybe from older version more prone, maybe something different and similar in the ones that need kicked?

[Hightech]

I can confirm v1.5.25 has fixed my SFP port issue.
But i now have a strange issue with "Netonix Manager v1.0.20" the table is mixed up and it only shows switches until the first netonix switch with v1.5.25 firmware, have tryed to reboot but it dident help

So what we need here is some more info and feed back.

Did you see this issue with v1.5.24 or what version did this start?

If there are any other users using manager and v1.5.25 do they see this and info as to what version you were using (manager and switch)and what version caused it?

I have no issue on v1.5.24 its begin on v1.5.25.

As soon i delete switches with v1.5.25 from Netonix Manage, the Manager interface is ok again.
The Manager fails to load the tabel when some data it get from the switch is not correct at the tabel crash when refreshing information.
Another issue we have had for a long time is: we use long names so the names and model field text are written on top of each other, is it possible to make the name field longer or auto length?

when trying to open switch webinterface from the Netonix Manager the token is broken og the URL is incorrect this happend from the first rewrited firmware v1.5.17 so have to delete part of the url to login to the switch

[oeyre]

So to confirm, some (85%) switches when upgraded SNMP worked fine and some (15%) a reboot or toggle fixed it?? -- KINDA WEIRD???

If above is correct it would be helpful to "try" and figure out why, maybe from older version more prone, maybe something different and similar in the ones that need kicked?

That is the gist of it, I was not being particularly scientific about it or taking notes to see if there were any common factors. For this particular problem I have not seen it happen again past the initial ones that experienced it on upgrading.

We are currently running a limited test of 1.5.25 with a view to deployment in 2-3 weeks on our units with remaining issues (SFP & web login) and will advise if we see any further weird issues.

[KBrownConsulting]

Just upgraded my test WS-8-150-AC from 1.5.24 to 1.5.25 simply & successfully.

I can confirm the following items are working better:

- SMTP test when FromAddress empty now works - 1.5.25
- PingWatchdog alert GPS corruption fixed - 1.5.25
- Traceroute resolve ips checkbox fixed - 1.5.25

I also haven't noticed any erroneous alerts about the switch rebooting so it looks like the following is also good:

- false alert about the switch rebooting has been fixed - 1.5.25


But I still can't get non-watchdog email alerts to work.

As with the last couple versions the switch no longer seems to send Link Change alerts anymore. (I'm not sure exactly when it stopped working since I didn't catch it right away and haven't reloaded all the recent firmware releases to figure out which broke it, but I suspect it was around the time of major code rewrite in 1.5.17)

As I reported in 1.5.24, the switch is definitely recognizing a Link Change event took place as it's recorded in the log, but it simply doesn't try to send an email. (Nothing in the log about trying to send an email and one never shows up.)

Additionally in this version I decided to try and test the "Fans" alert and that one seemingly isn't working either. I'll admit I'm not 100% sure how the "Fans" alert is triggered, but "for science" I tried to test it by simply sticking something in the fan while it was spinning to bring it to a full stop. On the Status page the "Fan 1 Speed" changed to "ERR" but 30 seconds later it still hadn't tried to send an email...

The fact that I can't get either the "Link Change" or "Fans" alerts to work makes me suspect all 5 of the non watchdog email alerts are broken...

I do use the Link Change alerts and hope that can be fixed sooner rather than later and ironically, after using Netonix switches for over 5 years with no fan issues, in the last 6 months I've had fans die in 2 different switches and it was really nice to get an email alert about it...