v1.5.17 Bug Reports and Comments

[sirhc]

Bricked 2 WS-12-250-DC switches and your fix script didn't work.

Did you read, that fix script is for OLD units

OK so OLD switches:
WS-24-400
WS-24-400A
WS-12-250A
WS-8-250-DC (the old original ones with flat SFP stack)

May have a problem as in the bneginning we did not include boardrev in the config so they will have problems but they are 10 year old switches

There is a fix:
Connect console cable:
drop to linux with CMD command
Cut and paste this into console:

cmd
cat<<EOF>cfgfile
MAC=EC:13:BX:XX:XX:XX <= NOTE MAKE SURE MATCHES ORIGINAL MAC
----MODEL=WS-24-400
----MODEL=WS-24-400A <=USE WHAT YOUR MODEL IS
----MODEL=WS-12-250A

BOARDREV=X <=USE WHAT YOUR BOARD REV WAS BUT IF UNSURE USE "B"
SERIAL=0
EOF
mtd write cfgfile config
reboot [CR in paste or Enter at end]

A WS-12-250-DC is NOT an old unit

If the switches are power cycled during upgrade you can brick the unit. Some people use a smart power strip that will reboot a switch if pings stop. This upgrade process can take longer than normal due to a lot of house keeping from the vast changes in code. So if the unit is power cycled during the upgrade it is bricked. Make sure to disable this or increase time missing pings till reboot 10 minutes or more. I know this as I had this issue on one of my towers using a smart power strip. You might be bale to recover it via console cable in your lab there is a post on here that details how.

We upgraded many WS-12-250-DC units as well as EVERY unit we ever built with no issues.

Here are some post detailing this recovery process:
viewtopic.php?f=17&t=7999&p=38375&hilit=redboot#p38375

viewtopic.php?f=17&t=7999&p=38369&hilit=redboot#p38369

[Stephen]

Bricked 2 WS-12-250-DC switches and your fix script didn't work.

We will need more details to help.

I have re-run this upgrade from 1.5.14, 1.5.16, and 1.5.17rc2 on the WS-12-250-DC just to be sure something fundamental isn't wrong and it appears OK.

What firmware was it on? On our testing we went back as far as 1.4.9 to make sure upgrades would work, but coming from older versions like this can still be difficult for an older switch.

What was is the mac address for these 2 units?

If you are able, can you provide the config for the switch's? If not because access even to the console isn't working anymore, please try using the recovery tool mentioned above.

If possible try using a console cable and save the output of the bootup messages (as far as it is able to get at least) to a text file and upload it here for analysis.

As sirhc mentioned the script he provided should not be used casually, even on older units, if you where already on 1.5.17rc2 you would not need it and it is only required on the specific models he has in the post.

[softcombrian]

Bricked 2 WS-12-250-DC switches and your fix script didn't work.

Did you read, that fix script is for OLD units

Yes, I can read, no need to be hostile. The switches were bricked in the field after doing the upgrade, I didn't run your scripts until one of them was back here on the bench and was trying to get it to come alive.

A WS-12-250-DC is NOT an old unit

I'm not a Netonix expert, I have no idea what "OLD" means. I didn't even know what year it was made until I removed the cover off the unit and saw a copyright date.

If the switches are power cycled during upgrade you can brick the unit. Some people use a smart power strip that will reboot a switch if pings stop. This upgrade process can take longer than normal due to a lot of house keeping from the vast changes in code. So if the unit is power cycled during the upgrade it is bricked. Make sure to disable this or increase time missing pings till reboot 10 minutes or more. I know this as I had this issue on one of my towers using a smart power strip. You might be bale to recover it via console cable in your lab there is a post on here that details how. We upgraded many WS-12-250-DC units as well as EVERY unit we ever built with no issues. Here are some post detailing this recovery process: viewtopic.php?f=17&t=7999&p=38375&hilit=redboot#p38375 viewtopic.php?f=17&t=7999&p=38369&hilit=redboot#p38369

* The unit never lost power during the upgrade.
* No ping watchdog set on the switch or the power connected to the switch.
* Brought the unit back from the field. Ran the recovery process for 2+ hours, sat at 100%.

[softcombrian]

Bricked 2 WS-12-250-DC switches and your fix script didn't work.

We will need more details to help.

I have re-run this upgrade from 1.5.14, 1.5.16, and 1.5.17rc2 on the WS-12-250-DC just to be sure something fundamental isn't wrong and it appears OK.

What firmware was it on? On our testing we went back as far as 1.4.9 to make sure upgrades would work, but coming from older versions like this can still be difficult for an older switch.

What was is the mac address for these 2 units?

If you are able, can you provide the config for the switch's? If not because access even to the console isn't working anymore, please try using the recovery tool mentioned above.

If possible try using a console cable and save the output of the bootup messages (as far as it is able to get at least) to a text file and upload it here for analysis.

As Chris mentioned the script he provided should not be used casually, even on older units, if you where already on 1.5.17rc2 you would not need it and it is only required on the specific models he has in the post.

Switch was on 1.5.17rc2

MAC address for one is: EC:13:B2:85:AD:D2. The other one is being brought back from the field now.

Disconnecting the i2c cable to the DC part of the switch allowed it to boot and become responsive, at which point I cleared the config, rebooted, and then reflashed 1.5.17 onto it. Powered the switch down after it was completed, reconnected the i2c cable, and got this through the serial console log (loops this repeatedly):


Dec 31 16:00:15 netonix: 1.5.17 on WS-12-250-DC
Dec 31 16:00:15 sysinit: switch[695]: switch proces waiting for sync to settl
Dec 31 16:00:15 switch[695]: switch proces waiting for sync to settl
Dec 31 16:00:16 sysinit: switch[695]: switch proces waiting for sync to settl
Dec 31 16:00:16 switch[695]: switch proces waiting for sync to settl
Dec 31 16:00:17 sysinit: switch[695]: switch proces waiting for sync to settl
Dec 31 16:00:17 switch[695]: switch proces waiting for sync to settl
Dec 31 16:00:18 sysinit: switch[695]: switch proces waiting for sync to settl
Dec 31 16:00:18 switch[695]: switch proces waiting for sync to settl
Dec 31 16:00:19 sysinit: switch[695]: switch proces waiting for sync to settl
Dec 31 16:00:19 switch[695]: switch proces waiting for sync to settl
Dec 31 16:00:20 sysinit: switch[695]: switch proces waiting for sync to settl
Dec 31 16:00:20 switch[695]: switch proces waiting for sync to settl
Dec 31 16:00:21 sysinit: switch[695]: switch proces waiting for sync to settl
Dec 31 16:00:21 switch[695]: switch proces waiting for sync to settl

Also, there's a problem in your NTP script:

Dec 31 16:00:06 sysinit: /sbin/hotplug-call: /etc/hotplug.d/iface/20-ntpclient: line 3: syntax error: unexpected "("

[Stephen]

Dec 31 16:00:15 netonix: 1.5.17 on WS-12-250-DC
Dec 31 16:00:15 sysinit: switch[695]: switch proces waiting for sync to settl
Dec 31 16:00:15 switch[695]: switch proces waiting for sync to settl
Dec 31 16:00:16 sysinit: switch[695]: switch proces waiting for sync to settl
Dec 31 16:00:16 switch[695]: switch proces waiting for sync to settl
Dec 31 16:00:17 sysinit: switch[695]: switch proces waiting for sync to settl
Dec 31 16:00:17 switch[695]: switch proces waiting for sync to settl
Dec 31 16:00:18 sysinit: switch[695]: switch proces waiting for sync to settl
Dec 31 16:00:18 switch[695]: switch proces waiting for sync to settl
Dec 31 16:00:19 sysinit: switch[695]: switch proces waiting for sync to settl
Dec 31 16:00:19 switch[695]: switch proces waiting for sync to settl
Dec 31 16:00:20 sysinit: switch[695]: switch proces waiting for sync to settl
Dec 31 16:00:20 switch[695]: switch proces waiting for sync to settl
Dec 31 16:00:21 sysinit: switch[695]: switch proces waiting for sync to settl
Dec 31 16:00:21 switch[695]: switch proces waiting for sync to settl

Also, there's a problem in your NTP script:

Dec 31 16:00:06 sysinit: /sbin/hotplug-call: /etc/hotplug.d/iface/20-ntpclient: line 3: syntax error: unexpected "("

That's interesting, please try re-downloading 1.5.17 and let me know how if the behavior changes. And delete the 1.5.17 you previously downloaded.

[sirhc]

softcombrian sorry if I seemed gruff, I'm extremely tired and burnt out.

Please follow Stephen's instructions and it should fix your problem.

Just make sure you re-download v1.5.17 and delete previous version and you should be fine.

Sorry for the extra work and please let us know if it works for you.

[Banana Jack]

Thanks for the new firmware. I upgraded three in-house units from 1.5.14 to 1.5.17 and all went well. The units were WS-12-250-AC, WS-6-MINI and WS-26-400-AC. The upgrade time was about 8 minutes and their connected devices lost LAN only for about a minute during the reboot. PoE was maintained throughout.

Netonix Manager didn't update its screen to show any of the the new versions (one of them was stuck showing "Rebooting"). I unticked "Monitor" and then ticked it again and this updated their shown versions and uptimes etc.

I plan to run these units for a month or two and if no problems then will slowly roll out to the full network, ~100 units.

Thanks again
Glenn

[Banana Jack]

If it weren't for the fact that I was upgrading a switch that a tech was in the process of installing, I'd have had a near heart attack. Upgrade itself went fine, got logged back in with admin/admin. Changed the password and logged out. Can't log back in. Try again. Fat finger? Copy/paste. Nope, it is not working. SSH? Yup, that works fine. Try resetting password again. Same results. Can't log in via Web, but SSH works fine. Having to use a shorter/simpler password (no special characters) for the moment.

This happened to me too. Can't log back in after putting a bang in the password. Do I have to factory reset or is there an easier way of re-gaining access?

[sirhc]

You really should be able to set password with console but make sure you hit enter after command in console after credential command to confirm config change.

I guess that required enter to commit change could be removed.

Some people report they can SSH into switch with default password after upgrade but not UI.

I'm thinking there are some cashed pages in browser? Try refreshing or clearing cache for address of switch.

I've switched to using Firefox for awhile as it easily allowed me set min TLS to 1 and its behavior is better IMO.

[sirhc]

Also those having no issues please report as 1 in 100 go out of way if things work good so you only hear those with issues which deters others.

Anyone who downloaded early should delete file and re-download as we suspect a corrupt file that effected ws-12-250-DC and possibly WS-26-500-DC UNITS.

REMEMBER AFTER UPGRADE IF YOU CAN'T LOGIN VIA UI WITH DEFAULT admin PASSWORD BUT CAN LOGIN TO SSH EITHER TRY REFRESHING OR DELETING BROWSER CACHE OR SETTING PASSOWRD IN SSH MAKING SURE TO HIT ENTER IN CONSOLE AFTER CREDENTIAL COMMAND TO CONFIRM CONFIG WRITE.

This really is BEST firmware since v1.4.8 if not best ever. There are quirks with forced password change due to upgraded hash storage of password for better security. And re write of login part of UI may cause issue from old page being cached.

I did see it once in chrome but not in Firefox.

In hindsight should have changed page name?