BBAI Cannot connect via web browser after build - A little help needed
Ok, so followed the alpha build instructions including doing the various BBAI updates first. All went as described, no error messages and compiled OK via various reboots. I can SSH into the BBAI with acct debian and password temppwd. if I ct (once I've done sudo bash), temp is about 41 degrees c with no activity.
However, when I try to connect via the web url http://192.168.0.14:8073/ , I get a response of;
192.168.0.14 refused to connect
what do I look for where or what do I check via the ssh to see what is happening. I think this is may be a pass wording or account issue? I imagine I need to dump a log somewhere and post it here but I'm no unix type so need some guidance please?
Has the account to log in via the web not created maybe because I did the build without the SDR plugged into the BBAI (something to do with serial numbers?)
I think it builds OK but as part of the start up looks for a valid FPGA/serial so won't actually run the web interface without a cape.
it probably shows something in /var/log/messages so "tail -f /var/log/messages" will give an updated visual report (or just "cat" it).
So now, I've done a make clean, make, make install again with the sdr plugged into the BBAI. Same state afterwards. All processed OK, no error messages. Rebooted. Just the same. Once additional point, If I go to port 80 I get the cloud 9 IDE so I know the device is accessible over the web. If I stop this service as per one of the other threads, still nothing on 8073
This is what I get with that tail command...
Apr 7 12:44:49 beaglebone kiwid: 00:00:00.499 compiled: Apr 7 2021 12:07:08 on beaglebone
Apr 7 12:44:49 beaglebone kiwid: 00:00:00.500 -debian 9
Apr 7 12:44:49 beaglebone kiwid: 00:00:00.501 /etc/debian_version 9.13
Apr 7 12:44:49 beaglebone kiwid: 00:00:00.502 background mode: delaying start 30 secs...
Apr 7 12:45:19 beaglebone kiwid: 00:00:30.538 reading configuration from file /root/kiwi.config/kiwi.json: 49 tokens
Apr 7 12:45:19 beaglebone kiwid: 00:00:30.539 reading configuration from file /root/kiwi.config/admin.json: 25 tokens
Apr 7 12:45:19 beaglebone kiwid: 00:00:30.540 EEPROM check: open /sys/bus/i2c/devices/3-0054/eeprom No such file or directory
Apr 7 12:45:19 beaglebone kiwid: 00:00:30.541 can't read serial number from EEPROM and no configuration override
Apr 7 12:45:19 beaglebone kiwid: 00:00:30.559 reading configuration from file /root/kiwi.config/dx.json: 7172 tokens
Apr 7 12:45:19 beaglebone kiwid: 00:00:30.559 898 dx entries
Apr 7 12:45:19 beaglebone kiwid: 00:00:30.564 .... firmware: SDR_RX4_WF4
Apr 7 12:45:19 beaglebone kiwid: 00:00:30.565 .... firmware: rx_chans=4 wf_chans=4
Apr 7 12:45:19 beaglebone kiwid: 00:00:30.566 .... firmware: RX bufs=4 samps=170 loop=85 rem=0 intr_usec=14166
Apr 7 12:45:19 beaglebone kiwid: 00:00:30.566 .... firmware: WF xfer=9 samps=911 rpt=50 loop=18 rem=11
Apr 7 12:45:19 beaglebone kiwid: 00:00:30.567 .... listening on default port 8073/8073 for "openwebrx"
Apr 7 12:45:19 beaglebone kiwid: 00:00:30.568 .... webserver for "openwebrx" on port [::]:8073
Apr 7 12:45:20 beaglebone kiwid: 00:00:30.809 .... ### using SPI_DEV
Apr 7 12:45:32 beaglebone kiwid: 00:00:00.473 KiwiSDR v1.450 --------------------------------------------------------------------
Apr 7 12:45:32 beaglebone kiwid: 00:00:00.475 compiled: Apr 7 2021 12:07:08 on beaglebone
Apr 7 12:45:32 beaglebone kiwid: 00:00:00.476 -debian 9
Apr 7 12:45:32 beaglebone kiwid: 00:00:00.477 /etc/debian_version 9.13
Apr 7 12:45:32 beaglebone kiwid: 00:00:00.478 background mode: delaying start 30 secs...
Apr 7 12:46:02 beaglebone kiwid: 00:00:30.514 reading configuration from file /root/kiwi.config/kiwi.json: 49 tokens
Apr 7 12:46:02 beaglebone kiwid: 00:00:30.515 reading configuration from file /root/kiwi.config/admin.json: 25 tokens
Apr 7 12:46:02 beaglebone kiwid: 00:00:30.516 EEPROM check: open /sys/bus/i2c/devices/3-0054/eeprom No such file or directory
Apr 7 12:46:02 beaglebone kiwid: 00:00:30.517 can't read serial number from EEPROM and no configuration override
Apr 7 12:46:02 beaglebone kiwid: 00:00:30.535 reading configuration from file /root/kiwi.config/dx.json: 7172 tokens
Apr 7 12:46:02 beaglebone kiwid: 00:00:30.536 898 dx entries
Apr 7 12:46:02 beaglebone kiwid: 00:00:30.540 .... firmware: SDR_RX4_WF4
Apr 7 12:46:02 beaglebone kiwid: 00:00:30.541 .... firmware: rx_chans=4 wf_chans=4
Apr 7 12:46:02 beaglebone kiwid: 00:00:30.542 .... firmware: RX bufs=4 samps=170 loop=85 rem=0 intr_usec=14166
Apr 7 12:46:02 beaglebone kiwid: 00:00:30.543 .... firmware: WF xfer=9 samps=911 rpt=50 loop=18 rem=11
Apr 7 12:46:02 beaglebone kiwid: 00:00:30.544 .... listening on default port 8073/8073 for "openwebrx"
Apr 7 12:46:02 beaglebone kiwid: 00:00:30.545 .... webserver for "openwebrx" on port [::]:8073
Apr 7 12:46:03 beaglebone kiwid: 00:00:30.786 .... ### using SPI_DEV
At the risk of making John scroll too much the "can't read serial number from EEPROM" means something is stopping the software or BBAI seeing the cape.
If you are sure you have plugged it in properly then confirm for me how you are powering the cape plus AI? (via the Cape barrel plug?)
yes, via the cape barrel plug with a 4A 5v PSU
that I was using on the BBG
Just to be sure of the PSU, I've hooked up a bench PSU set at 5.2V and max of 6A, The BBAI/SDR combo peaks up about 2.1A during start-up and settles at 1.3A no activity. Plugged it back into BBG to make sure I've not bricked it. No, still working fine in the BBG
I doubt you have bricked your SDR (listen if I still have my original working most people should be fine)
Wait for JKS, or if the main man is busy, I'll have a look at my AI when home*, best not to turn this into a long thread with guesses. (*I think better with a device in front of me).
First off, this is a genuine KiwiSDR board and not one of the recent clones, right?
Two things: What prints after the
### using SPI_DEVline? Or does it go into a restart loop at that point? At 12:45:20 it looks like it restarted, but I don't know if that's because you did something manually or not. After
### using SPI_DEVis when it first tries to access the FPGA.
Are you absolutely sure you plugged in the Kiwi board correctly into the Beagle headers? If you were "off by one" row, and powered up, the Kiwi board would be instantly destroyed as power gets applied to the wrong pins.
One quick thought
Could you see what the result is of the command
Yes it's genuine, bought from Martin Lynch &S or Farnell here in the UK in about 2016, has your name printed on the PCB John. I can send a photo of the serial number hand written in black pen on a white background if you like.
It must go into a restart loop at that point as I don't ever see anything past that point. Just see that loop over and over.
Yes, I'm sure it was plugged in correctly, very careful to check it before powering on. I've since plugged it back into the BBG to check I'd not damaged it and it's working fine in there right now. As you say, if I'd plugged it in wrong, it would be dead now. My call sign is G8PFR and it is now back on-line plugged into the original BBG. I'll plug it back into the BBAI tomorrow and carry out any instructions you give (or maybe you could log into it remotely if I opened up some ports on my router?). Any help you can offer is much appreciated. I'll also check the ic2 bus is seen tomorrow.
Since it's publicly listed I was able to connect. Seems fine running on the BBG. My guess is something has gone wrong with accessing BBAI devices from user space (I2C to EEPROM, SPI to Kiwi).
Not being able to read the serial number is not such a big deal. But not being able to talk to the FPGA is obviously. But there would have been a message to that effect. So that's worrying.
If you can open ssh port 22 on your router, and switch the Kiwi back to the BBAI, then I can login and try and see what's going on. It's possible that some recent BBAI update has broken things relative to the Kiwi distro. Anyone else do a BBAI update recently?
OK John, many thanks. I'll open up port 22 tomorrow and route it to the BBAI. Is there a way I can advise you of the log-in details a bit more privately?
Just for completeness
root@beaglebone:/home/debian# i2cdetect -l
i2c-3 i2c OMAP I2C adapter I2C adapter
i2c-0 i2c OMAP I2C adapter I2C adapter
Ready for you John as per email
Just for info jks (ZL) is about 11hr's ahead of us so response is more likely in his normal daylight hours. (nearly midnight there right now). Also wroth checking this forum early in the morning UK time.
Thanks for the i2cdetect output, looks similar to what I saw when mine was on last evening.
Well, I don't know. It looks like a recent BBAI update has screwed everything up. Not to mention that someone is closing my ssh into your machine randomly every 15 - 45 seconds after I connect. Extremely annoying.
I'm going to have to put this one in the queue. It's simply not important at the moment. My inbox is full again. I've got lots of people complaining that the proxy isn't working for them. I've got other problems.
Not sure why the ssh is being closed. I've just routed it straight through ( I hope). Should I leave it set up this way for a few days in case you get a change to look at it again or will you take a look on a test system of your own? regards Mike
The SSH thing is probably the ISP, in the UK the most common IPS's sell the "benefit" of a "clever" firewall that does that unless we sign wavers and forgo the hand holding.
@Gibsonmb you probably can adjust the ISP firewall settings and rely on your router firewall.
The clones should not benefit directly from your work, i know its a license thing but there should be a cut off point, they made tweaked their own hardware and should inherit the software version up to the point that became a commercial item. Mind you first clone complaining about the second clone "copying their work and only making small changes" was worth a chuckle.
I'm guessing this is an SPI overlay issue and going back to an earlier kernel might be a work around. I know the PI based ones changed overlay and I saw this sort of thing.
I only have the router from BT. It has a firewall build in but once I've set a port open, I can normally log in without problem from other locations. The firewall can be turned off but you still have to put the port routing in. I can of course set the external port to anything pretty much and then have the router translate it to 22 internally if that helps at all.
You should immediately change the debian account password (if the Kiwi software had gotten further it should have done that automatically) and email it to me. There are bots that scan the web looking for open ssh ports with default passwords, including on Beagles, then implanting viruses. We've actually seen this.
@Stu: The mentality coming out of Asia currently (and elsewhere too, let's not kid ourselves) will be the death of open source and, to a lesser degree, open innovation in my opinion.
I haven't done any searching on the web to see if this has been discussed, but it almost certainly has. Now strictly speaking in general I'm not against what's happening. It's represents disruptive change which has always benefited technology. It's hurting me personally, but that's just my narrow, short-term view related to the Kiwi. There's nothing wrong with anyone wanting to buy cheap, as long as they're prepared to deal with the consequences (questionable support, etc).
I've had ideas for follow-on Kiwi products for a long time now and those plans are completely dead at this point. Even with new closed-source software. It's just not worth what remaining lifetime I have left. There are better things to be doing..
@jks I hear you on opensource trust/use and have see the 1:1.2 hardware cloning thing first hand elsewhere, the one good thing is the usual risible support for a clone, so commercial customers know that it is not about just purchase price and a couple of extra features, it is about the life and support of the item. That's what stings here, the things they don't do well after hardware, support and upgrades you are picking up, seemingly at 4:30am - midnight.
I'll have a dabble with my AI as I've been meaning to try a more recent image for a while (reports of less heat) if I can find a 1-2-3 route with a newer fresh image I'll post that so I don't risk someone esle's hardware direct (Numpty is in the username for a reason).
Just for the record still wish I'd found the Kiwi sooner, best ever radio purchase.
I agree on the quality of this radio and also, not supporting low value, low support clones. I bought this in 2016 and it's still going strong. I bought it as a complete kit with the BBG and added the SEEED Studio aluminium case. I've got an SD card at the ready with the latest BBAI debian downloaded and ready to reinstall should that be worth a try. The reason I've not tried that first is because with my limited capabilities I had an awful job of disabling the DNS on it trying to use the USB eth route rather than the wired RJ45 ethernet as default... oh and forcing a static ip address, which is probably really simple but connmanctl was just not working for me so took me ages just to set simple basics up.
Okay, all should be good now (except for my splitting headache). I'm not 100% sure because it won't let me stay logged in for more than a few seconds (BT bastards). And your router isn't forwarding port 8073 connections to the new local ip address of course.
Sure enough, the Beagle guys made an incompatible change in how the .dtb file is loaded by Uboot. I'll have to figure out a change to support it.
Great, looks to be working now. I've powered it down short term because now of course it's running at 1.5Ghz and thus I need to tweak the cooling. I also need to do a bit work on the filtering on the linear PSU feeding it and move the sense leads closer to the BBAI... oh and swap the DC supply lead lead for heavier wire. Don't want those nasty little voltage drops arising at the BBAI input! Thank you very much for the work you've done and continue to do on this John. I should have it publicly on-line over the weekend. PS I don't even know what a dtb file is.
Back on-line in public now. Just waiting for larger end fan to arrive to supplement the Farnell cape fan as temp still a bit high. I will also raise the sdr away from the fan cape with header risers..... also awaited. Thanks again for your support