v1.4.0 Bug Reports and Comments

[sirhc]

Please post comments and report bugs with v1.4.0 FINAL in this thread

This version fixed a LOT of issues (see below) so we strongly suggested upgrading your switches.

If you try this version and still have issues try the latest RC version as this version is FINAL and no development is being done in it.

FIXED
- Fixed SNMP reported uptime
- Fixed rare problem with mac table not loading
- Fixed small UI issues with Firefox and IE
- Fixed SNMP CPU counters
- Fixed SNMP firmware version string
- Fixed so log messages now used VLAN ID instead of index
- Fixed RADIUS to work with SSH
- Fixed port not being in VLAN Bug
- Fixed pps being included in totals when port was excluded
- Fixed false fan failure detection and alerts
- Fixed cleared log to indicate who cleared it
- Fixed so no duplicate VLAN ID allowed
- Fixed SSH session ending messages so not reported by dropbear
- Fixed Firefox auto complete (workaround)
- Fixed console in UI
- Fixed graphs on Status Tab
- Fixed some incorrect voltage readings/alerts
- Fixed I2C communications problems mostly for SMART DC models
- Fixed RSTP path cost now able to set to 1
- Fixed Memory leak
- Fixed log messages to not report successful logins (dropbear)
- Fixed 10MB led
- Fixed IGMP Snooping issue
- Fixed utility packets using Tx Q1 instead of Q0

ENHANCEMENTS
- Increased MAX allowed DC input voltage for WS-8-250-DC and WS-12-250-DC to 61.5V for off grid solar sites
- IMPROVED SMART DC switches I2C communication
- Added HTTP Port to be changed
- Added telnet for CLI access
- Added Discovery to Tools menu
- Added multiple languages support
- Added VLAN information to SNMP
- Added Revert Timer Help Balloon
- Added total power consumption to SNMP
- Added DC input voltage graph on the Device/Status Tab
- Added tool tips to some buttons
- Added increased SMTP Username length
- Added Power tab on all models for time conditions
- Change ifDescr and ifAlias
- Time Condition graphic on Christmas tree
- Added To line to email header
- Added more graphs for SMART DC models under Device/Status Tab
- Added an Enable/Disable check box for TFTP backup
- Reduced CPU usage
- Added Flow Control Pause Frame Storm Protection
- Added the ability to configure Flow Control Obey/Generate/Both per port

KNOWN ISSUES
- Some language templates need help
- Discovery interfered with STP and LACP and caused high CPU utilization
- Could not disable Loop Protection
- UI Formatting bugs

Released 5/20/2016

[sirhc]

I updated all my WISP switches to v1.4.0 FINAL

WORKS GREAT

SMART DC models check out the new graphs on the Device/Status Tab

Check out the Tool/Discovery Tab - AWESOME
If anyone has problems this can be disabled on the Device/Configuration Tab
We did not have the ability to test LARGE FLAT network segments

Those who do not like English try the Language Template Device/Configuration Tab
WE KNOW SOME LANGUAGES NEED HELP SO IF YOU WANT TO HELP EMAIL ME

[PBX-Tech]

Had an issue upgrading from 1.4.0rc25 through Netonix Manager v1.0.0rc6. It would download the file then just sit at Uploading and never budge. I ended up just downloading the 1.4.0 file locally and then logged into the switch and upgraded via Utilities/Upgrade without any problem.

Here is the log from Netonix Manager.
5/20/2016 8:44:01 PM Upgrading device 10.19.70.11 from 1.4.0rc25 with firmware wispswitch-1.4.0.bin
5/20/2016 8:44:02 PM Error in POST /api/upgradefirmware: Error: ENOENT: no such file or directory, stat 'undefined/firmwares/wispswitch-1.4.0.bin'

[sirhc]

Manager worked for me?

This seems like a Manager issue not a firmware issue, please post about it in the Manager Thread:
viewtopic.php?f=6&t=1549

I would provide Eric with as much detail about your Manage Install as possible.

[PBX-Tech]

Will do tomorrow. Thanks.

[wisp-wand]

Hi Chris, as you may remember I was having issues with my site that appears to have high emc/interference (but NOT a FM site) I was running RC22 very well, and it was looking good. I upgraded to RC25 and the switch crashed again after less than 24hours. Basically nothing logged, its dead (apart from LED you can see lit inside) and needs a power cycle to recover it.
I have gone back to RC22 it has been up for 1 day 12 hours so looks ?ok?. I still see spikes when I look in the status page, for example Board temp jumps to 255 then back to normal and fan then goes to max, but other than that the switch does not react, in that it doesn't shut any ports down. Should I move to 1.4 final or stay on RC22?

Regards

Wayne

[cbl]

Upgraded a few devices to 1.4.0...couple of issues.

#1. One mini shortly after upgrading decided to stop responding to SNMP. An snmpwalk on the IP would result in "Timeout: No response from <ip>". Changing the version to 1 on the switch, would allow a -v1 snmpwalk to complete, and then changing back to v2c on the switch would allow a v2c snmpwalk to complete again.

#2. On one Ws-24-2400A we have a bunch of UBNT Titaniums. We've experienced the port lockup on these when set to Auto and have forced to 100FD on both the switch and AP to get around the AP deciding to become unreachable. One AP on this switch after update (from 1.3.9) was showing Tx Drops/Collisions incrementing, whereas before the upgrade it was not.

#3. On these 100FD forced ports, it looks like flow control icons on the Status page were displaying. They were not prior to 1.4.0 as from what I remember reading from the spec you can't have flow control with a non-Auto port. Just wondering if it's perhaps showing the icon regardless of if it's negotiated flow control with the other side?

#4. Random weirdness. On a WS-12-400-AC (was on 1.3.9rcSomething) that we have powering a MetroLinq 60GHz, the radio and devices behind it were un-reachable after update. Port bounce brought it back up. More than likely it's a MetroLinq problem, but thought I'd report it.

Loving 1.4.0 so far!!

[6gnet]

I still have issues with IGMP which simply doesn't work with tagged VLANs

-G

[sirhc]

Hi Chris, as you may remember I was having issues with my site that appears to have high emc/interference (but NOT a FM site) I was running RC22 very well, and it was looking good. I upgraded to RC25 and the switch crashed again after less than 24hours. Basically nothing logged, its dead (apart from LED you can see lit inside) and needs a power cycle to recover it.
I have gone back to RC22 it has been up for 1 day 12 hours so looks ?ok?. I still see spikes when I look in the status page, for example Board temp jumps to 255 then back to normal and fan then goes to max, but other than that the switch does not react, in that it doesn't shut any ports down. Should I move to 1.4 final or stay on RC22?

This particular site does appears to have some type of emc/interference that messes with the I2C communications and software is not going to fix this.

Options:
1) Use an AC model of the WISP Switch as theu do not suffer from this issue and use this DC switch somewhere else.

2) Track down the source of the interference

3) Wait for the new model WS-8-150-DC or wait for the new Board RevF of the WS-12-250-DC as these models are better suited to deal with interference in the FM band.

[sirhc]

#1. One mini shortly after upgrading decided to stop responding to SNMP. An snmpwalk on the IP would result in "Timeout: No response from <ip>". Changing the version to 1 on the switch, would allow a -v1 snmpwalk to complete, and then changing back to v2c on the switch would allow a v2c snmpwalk to complete again.

I do not know, maybe this was an anomaly? See if you can re-create it.

#2. On one Ws-24-2400A we have a bunch of UBNT Titaniums. We've experienced the port lockup on these when set to Auto and have forced to 100FD on both the switch and AP to get around the AP deciding to become unreachable. One AP on this switch after update (from 1.3.9) was showing Tx Drops/Collisions incrementing, whereas before the upgrade it was not.

I thought they fixed this issue in their firmware?
I think remember talking to several people that said they now use Auto again?
Try a port bounce to cause the Ethernet link on the radio to re-negotiate.
Nothing has changed that would affect Ethernet Communications as the core code has not been touched.

#3. On these 100FD forced ports, it looks like flow control icons on the Status page were displaying. They were not prior to 1.4.0 as from what I remember reading from the spec you can't have flow control with a non-Auto port. Just wondering if it's perhaps showing the icon regardless of if it's negotiated flow control with the other side?

Most devices will not negotiate Flow Control if not set to Auto some will. If we are indicating Flow Control is active then it is which you should be able to tell if Pause Frames are recorded under port stats.

#4. Random weirdness. On a WS-12-400-AC (was on 1.3.9rcSomething) that we have powering a MetroLinq 60GHz, the radio and devices behind it were un-reachable after update. Port bounce brought it back up. More than likely it's a MetroLinq problem, but thought I'd report it.

I got nothing.