/bin/false/.postgresql/postgresql.crt: not a directory

Setup and configuration
1

anyone came across the error?

Apr 30 09:28:29 instance-1 lora-app-server[2092]: time="2018-04-30T09:28:29Z" level=error msg="ping database error, will retry in 2s: stat /bin/false/.postgresql/postgresql.crt: not a directory"
Apr 30 09:28:31 instance-1 lora-app-server[2092]: time="2018-04-30T09:28:31Z" level=error msg="ping database error, will retry in 2s: stat /bin/false/.postgresql/postgresql.crt: not a directory"
Apr 30 09:28:33 instance-1 lora-app-server[2092]: time="2018-04-30T09:28:33Z" level=error msg="ping database error, will retry in 2s: stat /bin/false/.postgresql/postgresql.crt: not a directory"
Apr 30 09:28:35 instance-1 lora-app-server[2092]: time="2018-04-30T09:28:35Z" level=error msg="ping database error, will retry in 2s: stat /bin/false/.postgresql/postgresql.crt: not a directory"
Apr 30 09:28:37 instance-1 lora-app-server[2092]: time="2018-04-30T09:28:37Z" level=error msg="ping database error, will retry in 2s: stat /bin/false/.postgresql/postgresql.crt: not a directory"
Apr 30 09:28:39 instance-1 lora-app-server[2092]: time="2018-04-30T09:28:39Z" level=error msg="ping database error, will retry in 2s: stat /bin/false/.postgresql/postgresql.crt: not a directory"
2

Hello,
I have the same problem, after the update of an existing version of loraserver that was working very well.
I updated it today May 12th and since now, the loraserver service do not start with systemd (not being able to connect to the postgresql database with the same error message as above).

BUT, when I start the loraserver by simply typing “sudo loraserver”, then everything is perfect…

I don’t know what is happening, somebody has any clue ?

Thanks
Lucas

3

I have never experienced this error. Could you let me know how I can reproduce this so that I can look into this?

4

Hello Mr. Brocaar,
What we did was very simple : we were running a full Lora server for already 6 months (Ubuntu LTS16.04 Server) or so without any update (neither Ubuntu nor Loraserver) and, as we were starting a new database server, I decided, “well, why not go for a full update…” Bad, very bad idea :slight_smile:

So I just made the classic sudo apt-get update and sudo apt-get dist-upgrade.
I first noticed that the lora-app-server was not working any more, so I looked into the documentation and noticed that it was necessary to install an “add-on” to postgres. I did it and lora-app-server started working again. Then I noticed that the loraserver was not working so I checked the journalctl and saw that there was a postgres connection problem. I checked the .toml config file, the documentation, I logged into the database (loraserver_ns) manually, I changed the password (there was some !! inside the password, I thought it could be a problem). Nothing worked…

Then I decided to just type “sudo loraserver” and the server started flawlessly.
I moved the .toml file, tried “sudo loraserver --config-file /path/to/toml”. Again, flawless.
I tried again with systemctl : impossible to start the service properly : the service starts but journalctl shows the message on the first post.

I’m no Computer Hero to be honest, just very low to average user, but this time, I’m just clueless (systemd problem ? access right to .toml problem ?)…

The worst : I upgraded my own computer (Ubuntu 16.04LTS) and my version of the Loraserver suite is running flawlessly. I checked the postgres versions, they are the same on the server and my computer…

Lost in the high seas of bugs !

Thanks,
Lucas

5

The only think I can think of is that the dist-upgrade might have broken your configuration. On install, the LoRa (App) Server .deb installation script makes some assumptions about your environment, see:

https://github.com/brocaar/lora-app-server/blob/master/packaging/deb/post-install.sh

https://github.com/brocaar/loraserver/blob/master/packaging/deb/post-install.sh

Try completely removing loraserver and lora-app-server, including the related init.d and systemd files, /etc/default/loraserver and /etc/default/lora-app-server etc… You can keep the /etc/loraserver and /etc/lora-app-server configuration.

Btw, I have not had the time yet to test LoRa (App) Server on Ubuntu 18.04 LTS.

6

Dear Mr. Brocaar,
Ok, I will try the way you said and I will keep you informed.
We are on Ubuntu 16.04, sorry if I mistyped.
Sincerely yours,
Lucas

7

Hello Mr. Brocaar, hello everybody.
The problem was solved by:

  • removing entirely the loraserver suite (the 3 applications)
  • removing altogether mosquitto, redis and postgresql (after backup of the databases)
  • deleting all config files left behind (both from loraserver suite and from mosquitto)
  • updating the postgresql apt definition (as explained on the loraserver site)
  • updating the packages definitions
  • reinstalling mosquitto, redis and postgresql
  • upgrading the packages (but not dist-upgrade)
  • reconfiguring mosquitto
  • installing and configuring each elements of the loraserver suite one by one

Now, everything is working again. We are using Postgresql 10.4 version.

Whoaw, what a week-end ! Never ever a dist-upgrade again :slight_smile:

By the way, we still have problems with downlink messages, including OTAA, I will document it and complete the post I started on the subject

Cheers, have fun with LoraWAN and once again (never enough) thanks to Mr. Brocaar for this fantastic tool,

Luc

1 Like