Dietpi: Open Beta v6.27 | Please help testing and hardening the upcoming release

Created on 8 Dec 2019 · 35Comments · Source: MichaIng/DietPi

RC version | v6.27.2
---------- | -------
Changelog | https://github.com/Fourdee/DietPi/blob/beta/CHANGELOG.txt
Code changes | https://github.com/Fourdee/DietPi/compare/master...beta
v6.27.0 to v6.27.1 | #3270
v6.27.1 to v6.27.2 | #3282
How to apply | https://github.com/Fourdee/DietPi/blob/master/BRANCH_SYSTEM.md
Release planned | New Year 🎆

Important testing cases:

New Amiberry release, especially on Stretch systems: #3156
New Home Assistant with Python3.8: #3219
RPi4: EEPROM update implementation: #3217
New software title TasmoAdmin: #3103

Closes issues:

1615 #2593 #2897 #3103 #3156 #3174 #3175 #3177 #3180 #3181 #3182 #3187 #3191 #3194 #3203 #3207 #3208 #3213 #3217 #3219 #3221 #3242 #3244 #3246 #3255 #3259 #3261 #3263 #3264 #3271 #3272 #3273 #3275 #3284 #3291 #3293

Known issues

DietPi functionality

DietPi-Config | Enabling WiFi + Ethernet adapters, both on different subnets, breaks WiFi connection in some cases: #2103
- Current workaround is to disable the adapter that is not in use (not for internet connection).
- Otherwise a custom routing table is required.

SBC/device related

RPi | On TigerVNC virtual desktop, LXAppearance hangs on dbus-launch: #1791
Pine A64 | Bluetooth add-on module currently not supported: #528
Odroid XU4 | Kodi freezes shortly on video playback: #2584
Rock64 | 3.5mm A/V jack is currently not functional: #2522

Software title related

DietPi-Software | Node-RED: Pre-installed modules cannot be updated via web UI: #2073
DietPi-Software | Raspimjpeg: With Lighttpd, streaming mjpeg does not work: #1747
DietPi-Software | MATE desktop: When logging in as root, desktop items and right-clock context menu is missing: #3160
DietPi-Software | Sonarr/Mono: With current Mono version 6, import to a file system without UNIX permissions support (exFAT, FAT32/vfat and NTFS without "permissions" option) fails, regardless of user/umask mount options: #3179
DietPi-Software | Transmission: On Raspbian/Debian Stretch, RAM usage raises unlimited over time: #2413

Beta Information

Source

MichaIng

Most helpful comment

@dezren39
No plans to do so. We use the master branch for releases and do not want to keep older versions provided for download 😉. Also the raw code download does not make the system run, one needs to install it correctly and run the incremental patches, which is what dietpi-update does. So downloadable releases IMO give a wrong impression about these two things.

But yeah, I can imagine that following this project leads to a mass of notifications about commits and issue conversations. Actually I aimed to use PRs for everything, as well to have code changes, testing and changelog all done outside the dev branch and then merge it at once, which should reduce the amount of notifications/events on the commit-side. But especially for small changes with few code lines I am simply too lazy to do the PR overhead 😅.
The opened issues are again something out of our control of course.

Would be actually nice if it was possible to follow commits to a single branch only (Using master and beta branch then would keep you informed about all releases.), or excluding issue conversations and generally having some more control. Same in my case for Nextcloud and some other projects which are under HEAVY development, impossible to watch.

MichaIng on 21 Dec 2019

😄1 👍1

All 35 comments

Hi Guys,

for testing purposes, I was going to setup a DEV system. But I'm facing some issues during initial software installation. Could it be, that dietpi.com is offline??

───────────────────────────────
DietPi Error Handler: 
───────────────────────────────
 DietPi-Software: Connection test: https://dietpi.com/downloads/conf/desktop
- Exit code: 4
- DietPi version: v6.26.3 (MichaIng/master) | HW_MODEL:-1 | HW_ARCH:2 | DISTRO:5
- Image creator: DietPi Core Team
- Pre-image: Raspbian Lite

Log file content:
Spider mode enabled. Check if remote file exists.
--2019-12-09 23:31:13--  https://dietpi.com/downloads/conf/desktop
Resolving dietpi.com (dietpi.com)... 104.27.178.199, 104.27.179.199, 2606:4700:30::681b:b3c7, ...
Connecting to dietpi.com (dietpi.com)|104.27.178.199|:443... connected.
HTTP request sent, awaiting response... Read error (Success.) in headers.
Giving up.

Trying the website as well, is giving some Cloudflare's Always Online™ information, instead of the usual website :(

Joulinar on 9 Dec 2019

@Joulinar
I noticed this as well some hours ago, but when I was home it worked again. No hint on server logs, probably temporary VPS provider network issue.

Please try again now.

MichaIng on 10 Dec 2019

@MichaIng
hmmm I tried to install a VNC Server now, but it still keeps failing

─────────────────────────────────────────────────────
 Mode: Installing LXDE: ultra lightweight desktop

[FAILED] DietPi-Software | Connection test: https://dietpi.com/downloads/conf/desktop

After a Chromium test installation, it still fails to connect ssh.dietpi.com to send survey data.

[FAILED] DietPi-Survey | Connection test: ssh.dietpi.com

Joulinar on 10 Dec 2019

@Joulinar
Okay even stranger, since ssh.dietpi.com works from my module phone... I though it is related to Cloudflare, but seems not.

Works fine from my server as well:

2019-12-10 00:44:53 root@micha:/tmp# dietpi-survey 1
[  OK  ] DietPi-Survey | Connection test: ssh.dietpi.com
[  OK  ] DietPi-Survey | Successfully sent survey data

Server shows no overload, no error logs, nor anything else suspicious. Still VPS provider network issue, I guess. I have to sleep now, will check back tomorrow.

MichaIng on 10 Dec 2019

@MichaIng
Maybe it's depending on Geo Location. Will check it tomorrow as well. Good Night.

Joulinar on 10 Dec 2019

@MichaIng

Great work as always, thank you :+1:

Twitter REF: https://twitter.com/DietPi_/status/1204297321820934144

Fourdee on 10 Dec 2019

hmm still strange. I'm still getting the following once trying to open dietpi.com website

Unbenannt

Joulinar on 10 Dec 2019

@Joulinar
Could you try to disable any adblock and data savings feature? In my case access is only refused when using Operas data savings feature on mobile phone, which relays requests through Operas server then.

I tested with all firewall unblocked and disabled, but it is not our server itself which refuses the requests, they do not even reach it. It must be the VPS network.

MichaIng on 10 Dec 2019

I get the same error page.

DutchFlash on 10 Dec 2019

Lets move to #3260 with this topic...

MichaIng on 10 Dec 2019

@MichaIng
I was running a basic update on an empty system from v6.26.3 > v6.27.0 to test the RPi4: EEPROM feature. But I'm not sure if it was installed or activated. because I don't see anything within the logs. Or do I misunderstand how EEPROM is working? Within journalctl I don't see that much as well.

update_log.txt

Joulinar on 13 Dec 2019

@Joulinar
Many thanks for testing. Hmm is your RPi4 correctly recognised as such: echo $G_HW_MODEL

MichaIng on 13 Dec 2019

@MichaIng
hmm I guess that not the correct output :)

root@DietPi4:~# echo $G_HW_MODEL
-1
root@DietPi4:~#

Joulinar on 13 Dec 2019

@Joulinar
Okay, something wrong with model estimation, probably the syntax of the file which holds the revision code has changed. Now I remember a similar reported issue. Could you paste: cat /proc/cpuinfo

MichaIng on 13 Dec 2019

@MichaIng
see attached

cpu_info.txt

Joulinar on 13 Dec 2019

👍1

@Joulinar
Ah your board is a new revision PCB1.2. I'll have to add this to detection, at best switch to a generic code decoding method, since every bit has a meaning stating all the info up to RAM size.

MichaIng on 13 Dec 2019

@MichaIng
lucky me having a newer model :)

Btw I noticed something more after I switched to v6.27.0. During boot I'm getting the following lines.
20191213_123701

these are the corresponding lines from journalctl

Dez 13 12:33:36 DietPi4 systemd[1]: Starting DietPi-Boot...
Dez 13 12:33:36 DietPi4 systemd[1]: Starting /etc/rc.local Compatibility...
Dez 13 12:33:36 DietPi4 systemd[1]: Starting Permit User Sessions...
Dez 13 12:33:36 DietPi4 systemd[1]: Started /etc/rc.local Compatibility.
Dez 13 12:33:36 DietPi4 systemd[1]: Started Permit User Sessions.
Dez 13 12:33:36 DietPi4 sh[386]: eth0=eth0
Dez 13 12:33:37 DietPi4 systemd[1]: Starting Network Time Synchronization...
Dez 13 12:33:37 DietPi4 systemd[1]: Started Network Time Synchronization.
Dez 13 12:33:37 DietPi4 systemd[1]: Reached target System Time Synchronized.
Dez 13 12:33:37 DietPi4 kernel: bcmgenet fd580000.genet eth0: Link is Down
Dez 13 12:33:39 DietPi4 systemd[1]: systemd-rfkill.service: Succeeded.
Dez 13 12:33:40 DietPi4 kernel: bcmgenet fd580000.genet eth0: Link is Up - 1Gbps/Full - flow control rx/tx
Dez 13 12:33:40 DietPi4 kernel: IPv6: ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready
Dez 13 12:33:53 DietPi4 systemd-timesyncd[548]: Synchronized to time server for the first time 212.18.3.18:123 (0.debian.pool.ntp.org).
Dez 13 12:33:53 DietPi4 systemd[1]: Stopping Network Time Synchronization...
Dez 13 12:33:53 DietPi4 systemd[1]: systemd-timesyncd.service: Succeeded.
Dez 13 12:33:53 DietPi4 systemd[1]: Stopped Network Time Synchronization.
Dez 13 12:33:53 DietPi4 systemd[1]: Started DietPi-Boot.

Joulinar on 13 Dec 2019

👍1

@Joulinar
Okay I added support for this revision and possible further ones, at least when produces by Sony as well: https://github.com/MichaIng/DietPi/commit/b24939e16459f9489d8d0db5a9c9197fdb6b2f32
You can test:

wget https://raw.githubusercontent.com/MichaIng/DietPi/dev/dietpi/func/dietpi-obtain_hw_model -O /DietPi/dietpi/func/dietpi-obtain_hw_model
/DietPi/dietpi/func/dietpi-obtain_hw_model
. /DietPi/dietpi/func/dietpi-globals
echo $G_HW_MODEL # should now print "4"
/DietPi/dietpi/func/dietpi-set_hardware rpi-eeprom # Install EEPROM updater + update related firmware

MichaIng on 13 Dec 2019

🎉1

@MichaIng
jup, it's being detected correctly now.

root@DietPi4:~# echo $G_HW_MODEL
4

as well during login showing RPi4, which was not the case before.

 ─────────────────────────────────────────────────────
 DietPi v6.27.0 (beta) : 20:26 - Fr 13.12.2019
 ─────────────────────────────────────────────────────
 - Device model : RPi 4 Model B (armv7l)
 - Uptime : up 7 minutes
 - CPU temp : 40'C : 104'F (Optimal temperature)
 - LAN IP : 192.168.0.12 (eth0)
 - Info Text : !!! DEV Pi4 !!!
 ─────────────────────────────────────────────────────

and EEPROM is working finally.

*** INSTALLING EEPROM UPDATES ***
BOOTLOADER: update required
CURRENT: Mon 15 Jul 12:59:55 UTC 2019 (1563195595)
 LATEST: Tue 10 Sep 10:41:50 UTC 2019 (1568112110)
VL805: update required
CURRENT: 00013701
 LATEST: 000137ab
EEPROM updates pending. Please reboot to apply the update.
[  OK  ] rpi-eeprom  | Completed
root@DietPi4:~#

well done.

Joulinar on 13 Dec 2019

👍1

@Joulinar
Perfect, many thanks for testing 🚀.

MichaIng on 13 Dec 2019

Beta v6.27.1 has been merged: #3270

MichaIng on 21 Dec 2019

Hi! Are there any plans to use 'releases' in GitHub sometime in the future? I'm interested in keeping up with this project, but there is a _lot_ of activity day-to-day. I'd love if I could follow 'releases only' instead. 😅👍

dezren39 on 21 Dec 2019

MichaIng on 21 Dec 2019

😄1 👍1

@MichaIng
I was running a quick update check. For me it seems working fine without issues.

update.txt

Joulinar on 21 Dec 2019

❤1

@Joulinar
Great, many thanks for testing, also good to see EEPROM and asix ax88179 driver install going through nicely.

MichaIng on 21 Dec 2019

👍1

@MichaIng
Not sure where to place it but I notices some strange behavior. I was running Nextcloud for testing purposes and I removed it afterwards. However if I run dietpi-software > uninstall again, I still see MariaDB, Redis and Nextcloud even if it was already removed.

Unbenannt

Trying to uninstall them again will lead to an error message as the software is not available anymore :confused:

─────────────────────────────────────────────────────
 Mode: Uninstalling Nextcloud: File sync, sharing and collaboration platform

Ignoring unknown module: dietpi-nextcloud
Run "service lighttpd force-reload" to enable changes
grep: /var/www/nextcloud/config/config.php: No such file or directory
grep: /var/www/nextcloud/config/config.php: No such file or directory
grep: /var/www/nextcloud/config/config.php: No such file or directory
Failed to start mariadb.service: Unit mariadb.service not found.
/DietPi/dietpi/dietpi-software: line 12445: mysql: command not found
/DietPi/dietpi/dietpi-software: line 12449: mysqladmin: command not found
cp: cannot stat '/var/www/nextcloud/.': No such file or directory
rm: cannot remove '/var/www/nextcloud': No such file or directory

 DietPi-Software
─────────────────────────────────────────────────────
 Mode: Uninstalling Redis: optional non-sql database store

[ INFO ] DietPi-Software | APT removal for: redis-server redis-tools php*-redis, please wait...
[ INFO ] DietPi-Software | None of the requested packages are currently installed. Aborting...
[  OK  ] DietPi-Software | G_AGP redis-server redis-tools php*-redis

 DietPi-Software
─────────────────────────────────────────────────────
 Mode: Uninstalling MariaDB: database

[FAILED] DietPi-Software | systemctl start mariadb

Joulinar on 24 Dec 2019

@Joulinar
Strange, checked the code and couldn't find how this could happen. Running some test currently.
Do you have the end of the first uninstall attempt?

Please mark MariaDB manually as uninstalled:

G_CONFIG_INJECT 'aSOFTWARE_INSTALL_STATE\[88\]=' 'aSOFTWARE_INSTALL_STATE[88]=0' /DietPi/dietpi/.installed

We should probably remove the error handler for MariaDB start when its about to be uninstalled. This is to create a full database backup first, so users cannot accidentally remove important data.

EDIT: Just tested uninstall via GUI with Emby and it went as expected, hence the procedure generally works. It could only imagine that the uninstall process failed on your first attempt at some point, so that uninstall states were not written to disk anymore.

MichaIng on 24 Dec 2019

First of all I wish you Mary Christmas @MichaIng and many thanks for your incredible support over the last year. Much appreciated.

Coming back the the challenge 😄

I now removed LLMP, Redis and Nextcloud one by one again, using dietpi-software uninstall dialog. Even if nothing was done as no package exist, it's gone from the list. MariaDB is still failing as expected. However I created a Bug report, even if it is not needed.

reference code: 26a63a3e-1699-4250-8d78-67b8ba8155c9

Honestly I'm not sure what happen during the real first de-installation. It should have been removed sucessfull as I checked the software catalog and it was gone. I'm sure. Anyway I now set it back to 0 in /DietPi/dietpi/.installed as suggested.

Joulinar on 24 Dec 2019

👍1

@Joulinar
Since /DietPi/dietpi/.installed is on RAMdisk, it might be possible that some unclean shutdown or crash made the install state changes lost. This is one of the reason I want to move away from RAMdisk (#3288) for such state info, as well for the very important /DietPi/dietpi/.install_state e.g., which, when lost, would lead to firstrun setup and install steps run+prompt again on boot+login.

MichaIng on 29 Dec 2019

👍1

@MichaIng
Could be it was caused by an unclean shutdown as it happen on my test system. And usually I simply unplug power if not needed. :grimacing:

Joulinar on 29 Dec 2019

@MichaIng
I know it's spooky but MariaDB is back to my software catalog. Going to remove it again.

Unbenannt

Joulinar on 30 Dec 2019

@Joulinar
Indeed strange. Note that you need to do a regular shutdown now or reboot currently to have any changes to RAMdisk (including install flags) synced to disk. #3288 will allow to skip this by replacing the RAMdisk with a locked disk cache for DietPi files instead of copying and syncing back those to and from RAMdisk.

3294 on the other hand skips the error handled MariaDB start on uninstall if MariaDB was uninstalled previously already but uninstall flag has not been set or saved.

MichaIng on 31 Dec 2019

👍1

Honestly I never was taking care on these things. But starting now I will do a clean shutdown before going to deplug power. Btw, does poweroff is doing the same in regards to write back the RAMdisk?

Joulinar on 31 Dec 2019

@Joulinar
Jep, poweroff and shutdown now should do exactly the same.

MichaIng on 1 Jan 2020