No heartbeat or data after upgrade

I've done an upgrade to 232 and my Flukso has not transmitted data or had a heartbeat in 24 hours. I can see the device is connected to the network via Wi-Fi and I can access it via the wired LAN however the information page showing version number and device s/n is blank. Can anyone please assist in kick starting my device so it works again?
Regards,
Andrew

aennerst's picture

Really? No one has any advice for me?

gebhardm's picture

upg to v232, of course (how lovely it would be if one can edit own posts :-/)

aennerst's picture

Thanks for the reply. I have "Power Cycled" the device by pulling out the power lead. I don't know how to "reboot" other than this. I can see the mac address and allocated IP address on my router so it's connecting wirelessly. I can connect to it via LAN cable and can access it via 192.168.255.1 and I can see everything like normal however the first status box has nothing in it. I normally se the s/n and the firmware version (and other info I think but its raining at the moment so can't take the notebook out to check.
Is there a way to download the v232 firmware and then manually upload it? Settings/devices tab on flukso.net shows 232*** still after days so the upgrade has not taken and now its not responding at all.
Has anyone else had this problem??

Andrew

icarus75's picture

Can you try to get the FLM online again? I do not think any upgrade was initiated since it's still scheduled on flukso.net (232**). It seems like the FLM lost internet connectivity before the upgrade could even start. What does the syslog say?

aennerst's picture

Hi again,

I have plugged into the FLM and 100% it has internet connectivity. I can turn the wireless off on my notebook and connect to the net via the FLM. Device also has successful ping to flukso.net.
SYSTEM STATUS screen displays the following: Serial=Blank, Model=Blank, Hostname=Blank, Firmware Version=Blank, Time=Correct time and date displayed, Uptime=Correct uptime displayed.
As for the Syslog it's just repeating the same sequence every hour. (but none today for some reason)

Oct 25 06:00:05 flukso user.err heartbeat[10277]: WWW-Authenticate: No proper provisioning for this device
Oct 25 06:15:01 flukso cron.info crond[549]: crond: USER root pid 10312 cmd /usr/bin/fcheck
Oct 25 06:30:01 flukso cron.info crond[549]: crond: USER root pid 10342 cmd /usr/bin/fcheck
Oct 25 06:45:01 flukso cron.info crond[549]: crond: USER root pid 10372 cmd /usr/bin/fcheck
Oct 25 07:00:01 flukso cron.info crond[549]: crond: USER root pid 10402 cmd /usr/bin/heartbeat 0
Oct 25 07:00:01 flukso cron.info crond[549]: crond: USER root pid 10403 cmd /usr/bin/fcheck
Oct 25 07:00:05 flukso user.err heartbeat[10404]: POST https://api.flukso.net/device/0123456789abcdef0123456789abcdef: 401

I suspect the first line at 06:00:05 is the root of the problem but how can I solve it?

Regards,
Andrew

fury's picture

Mine has the same problem, I applied the update and at some point restarted the flukso. Now it just has the heartbeat error.

Jan 1 00:01:18 flukso user.info sysinit: sh: 1: unknown operand
Jan 1 00:01:18 flukso user.info sysinit: uci: Parse error (invalid command) at line 1, byte 0
Jan 1 00:01:18 flukso user.info sysinit: sh: 1: unknown operand
Jan 1 00:01:25 flukso user.err heartbeat[795]: POST https://api.flukso.net/device/0123456789abcdef0123456789abcdef: 401
Jan 1 00:01:25 flukso user.err heartbeat[795]: WWW-Authenticate: No proper provisioning for this device
Jan 1 00:01:26 flukso daemon.info supd[785]: Received event START for ntp

aennerst's picture

Yes FURY there seems to be a few people who are having an issue after upgrade. There is another person who has also started a thread with the same issue. It would be good too hear from someone in the know as to a possible solution or perhaps to roll back the firmware. It's frustrating that no one in development making any suggestions or asking for more information.

aennerst's picture

Is it possible to download the firmware and load it directly to the FLM??

sem's picture

according to your log file your time is off
try to change ntp setting

Jaap's picture

Same here

Firmware update showing as pending

Serial Version Resets Last heartbeat Uptime Operations
FL020xxxxx 232** 22 30 Oct 2013 12:44:20

  1. Jan  1 00:01:18 flukso user.info sysinit: uci: Parse error (invalid command) at line 1, byte 0
  2. Jan  1 00:01:18 flukso user.info sysinit: sh: FLM02B: unknown operand
  3. Jan  1 00:01:18 flukso user.info sysinit: sh: FLM02C: unknown operand
  4. Jan  1 00:01:18 flukso user.info sysinit: sh: 1: unknown operand
  5. Jan  1 00:01:19 flukso user.info sysinit: uci: Parse error (invalid command) at line 1, byte 0
  6. Jan  1 00:01:19 flukso user.info sysinit: sh: 1: unknown operand
  7. Jan  1 00:01:19 flukso daemon.info avahi-daemon[766]: Service "MQTT Broker on flukso" (/etc/avahi/services/mqtt.service) successfully established.
  8. Jan  1 00:01:20 flukso daemon.info avahi-daemon[766]: Service "Web Server on flukso" (/etc/avahi/services/http.service) successfully established.
  9. Jan  1 00:01:20 flukso daemon.info avahi-daemon[766]: Service "Flukso RESTful API on flukso" (/etc/avahi/services/flukso.service) successfully established.
  10. Jan  1 00:01:23 flukso user.err heartbeat[794]: POST <a href="https://api.flukso.net/device/0123456789abcdef0123456789abcdef:" title="https://api.flukso.net/device/0123456789abcdef0123456789abcdef:">https://api.flukso.net/device/0123456789abcdef0123456789abcdef:</a> 401
  11. Jan  1 00:01:23 flukso user.err heartbeat[794]: WWW-Authenticate: No proper provisioning for this device
  12. Jan  1 00:01:24 flukso daemon.info supd[784]: Received event START for ntp
  13. Oct 30 15:29:10 flukso cron.err crond[663]: time disparity of 23052447 minutes detected
  14. Oct 30 15:30:01 flukso cron.info crond[663]: crond: USER root pid 813 cmd /usr/bin/fcheck
  15. Oct 30 15:45:01 flukso cron.info crond[663]: crond: USER root pid 843 cmd /usr/bin/fcheck
  16. Oct 30 15:55:49 flukso daemon.info fluksod[790]: Caught a SIGTERM. Exiting...
  17. Oct 30 15:55:49 flukso daemon.info spid[787]: Caught a SIGTERM. Exiting...
  18. Oct 30 15:55:50 flukso daemon.info supd[784]: Caught a SIGTERM. Exiting...
  19. Oct 30 15:55:51 flukso daemon.info supd[931]: Sucessfully started with DEAMON=supd and DAEMON_PATH=/var/run/supd
  20. Oct 30 15:55:51 flukso daemon.info spid[934]: Sucessfully started with DEAMON=spid and DAEMON_PATH=/var/run/spid
  21. Oct 30 15:55:51 flukso daemon.info fluksod[937]: Sucessfully started with DEAMON=fluksod and DAEMON_PATH=/var/run/fluksod
  22. Oct 30 15:55:55 flukso user.err heartbeat[941]: POST <a href="https://api.flukso.net/device/0123456789abcdef0123456789abcdef:" title="https://api.flukso.net/device/0123456789abcdef0123456789abcdef:">https://api.flukso.net/device/0123456789abcdef0123456789abcdef:</a> 401
  23. Oct 30 15:55:55 flukso user.err heartbeat[941]: WWW-Authenticate: No proper provisioning for this device
  24. <code>

fury's picture

0123456789abcdef0123456789abcd is a genetic device code instead of a unuiqe one. So that's where the issue lies....

Jaap's picture

Yup /etc/config/system has this data I can enter the serial number but I have no ideas what device and key should be. I assume the provisioning mechanism shoudl initialize thsi after the first boot.

config system
option timezone UTC
option model FLM02A
option serial unknown
option hostname flukso
option device 0123456789abcdef0123456789abcdef
option key 00112233445566778899aabbccddeeff
option firstboot 1
option restore 1
option cronloglevel 1

sem's picture

serial number you can normally find on your dashboard on flukso.net

aennerst's picture

It's probably a dumb question but how do you get into /etc/config/system ?
I've tried putting this in the address line when connected directly to the FLM but it just kept taking me back to the root login.

Andrew

gebhardm's picture

Accessing the FLM's configuration on the level of the operation system requires a SSH client, e.g. PuTTY (see http://www.putty.org/ ) - with such a SSH client you can access the FLM via port 22; this brings you to the OS login (root, password root) and through that to the file system where you may alter nearly everything... (so you should know what you are doing - see OpenWRT documentation and the FLM's source code)
BUT for an upgrade that should not be necessary; so maybe Bart (the wizard behind the FLM) can help remotely; address him via the contact form to get a private conversation.

fury's picture

I don't know if that is very effective, for everyone to contact him. Surely they are already aware of the issue?

aennerst's picture

Should I be able to SSH the FLM via the wireless network it's connected too or do I need to be connected directly to the network port?

aennerst's picture

Got it. Thanks Gebhardm.

icarus75's picture

It seems that an upgrade to r232 was initiated. Then something seems to have gone wrong during first boot of the new firmware due to a UCI parsing error. That's why a.o. the device and key settings are still showing their defaults.
1/ Could you tell me which firmware version you upgraded from?
2/ If you're able to ssh into the FLM after a botched upgrade could you run this command:

  1. root@flukso-9212d0:~$ uci show system

I'm expecting the uci command to fail with the following error message:
  1. uci: Parse error (invalid command) at line 1, byte 0

Please copy/paste the /etc/config/system file into a comment as well.

And don't worry. We'll get these FLMs back online. You might have to reconfigure your FLM though.

Cheers
/Bart

sem's picture

ICARUS75 I had same problems.
Serial number was gone after reboot.
NTP settings were gone too . DNS was not working
root password changed also

icarus75's picture

@sem If your FLM is still in this state, could you please run the command in the above comment and let me know the result?

sem's picture

No, I fixed it already . But i need to reconfigure everything manually .
I just posting in case you are interested in improving upgrade process in the future.
If you need uci show system i can post it by mail to you

icarus75's picture

This issue seems to have popped up on a subset of the migrations. I haven't been able to reproduce the problem at my side. Investigating a broken upgrade can be very helpful to find the root cause.

Jaap's picture

  1.     OpenWRT AA [12.09.1, r38401]
  2.     Flukso [2.3.2-1, 173bda11c4f0c751]
  3.  
  4.   ------------------------------------------
  5.   Chuck Norris doesn't need a smart meter.
  6.   He just grabs the mains every hour to get
  7.   a feel for his electricity consumption.
  8.   ------------------------------------------
  9.  
  10. root@flukso:~# uci show system
  11. uci: Parse error (invalid command) at line 1, byte 0

Contents of /etc/config/system

(weird extra lines at start and end)

  1.  
  2. system                 0% |                               |     0  --:--:-- ETA
  3. config system
  4.         option timezone     UTC
  5.         option model        FLM02A
  6.         option serial       FL02000470
  7.         option hostname     flukso
  8.         option device       0123456789abcdef0123456789abcdef
  9.         option key          00112233445566778899aabbccddeeff
  10.         option firstboot    1
  11.         option restore      1
  12.         option cronloglevel 1
  13.  
  14. config timeserver ntp
  15.         list server     0.openwrt.pool.ntp.org
  16.         list server     1.openwrt.pool.ntp.org
  17.         list server     2.openwrt.pool.ntp.org
  18.         list server     3.openwrt.pool.ntp.org
  19.  
  20. config button
  21.         option button   reset
  22.         option action   released
  23.         option handler  net_toggle
  24.         option min      2
  25.         option max      5
  26.  
  27. config button
  28.         option button   reset
  29.         option action   released
  30.         option handler  net_defaults
  31.         option min      10
  32.         option max      30
  33.  
  34. config button
  35.         option button   reset
  36.         option action   released
  37.         option handler  firmware_defaults
  38.         option min      60
  39.         option max      120
  40.  
  41. config led led_wlan
  42.         option name     wlan
  43.         option sysfs    wlan
  44.         option trigger  netdev
  45.         option dev      wlan0
  46.         option mode     'link tx rx'
  47.  
  48. config counter event
  49.         option ath_kmod_reload  0
  50. system               100% |*******************************|  1002  --:--:-- ETA

Jaap's picture

And removing those first and last lines from /etc/system/config solves it

So:

Connect the flukso with a direct ethernet cable to your computer

Use putty (http://the.earth.li/~sgtatham/putty/latest/x86/putty.exe) or ssh to connect

  1. ssh 192.168.255.1

Login with username root password root

You can use vi to edit the file

  1.  vi /etc/config/system

Remove the first line by pressing the letter d twice

  1. dd

Go to the end of the file with the arrow keys

Remove the last line by pressing the letter d twice

  1. dd

Save the file by typing wq!

  1. :wq!

Reboot the router by typing reboot

  1. reboot

Jaap's picture

Please note do the above at your own risk maybe Bart has a better method

Only delete these lines if they show something similar to

system 0% | | 0 --:--:-- ETA

icarus75's picture

Allright! The migration script will download two files that will end up as /etc/config/system and /etc/config/firewall. I forgot the wget -q switch to silence the output of the wget command. The lines you're seeing is wget printing download progress. This has now been fixed in the script. For affected systems, the procedure given by jaap is correct, just make sure to scan both system and firewall, so:

  1. root@flukso-9212d0:~$ vi /etc/config/system
  2. root@flukso-9212d0:~$ vi /etc/config/firewall

If this seems too much trouble, then the alternative is to start the FLM, wait for the heartbeat LED to start blinking and then pressing the recessed toggle button between 60 and 120 secs. This will cause the FLM to delete the overlay file system containing the corrupted file(s) and to boot with a fresh r232 firmware. The downside is that your FLM will have lost all of its wifi and sensor settings. So you'll have to configure the FLM from scratch again.

@jaap Thanks for the help in finding the root cause. I've slightly edited two lines in your comment to fix typos.

aennerst's picture

Thanks JAPP and ICARUS75. removing the odd lines and reboot made the FLM bounce straight back into life. Still shows that the upgrade to 232 is pending so I'll see if it's successful this time. I'll report back in with my result.
Andrew