Open Beta v6.26 | Please help testing and hardening the upcoming release #3131
Comments
MichaIng
added
Testing/testers required 🔽
Information ℹ️
Beta 🧪
|
Currently running on Rpi3 . Keep getting error message on kodi installation.
`#### Details:
Steps to reproduce:
Expected behaviour:
Actual behaviour:
Extra details:
Additional logs:` |
|
I was running the update 2 times now in my HyperV VM. Both times I got the following error message Details:
Steps to reproduce:
Expected behaviour:
Actual behaviour:
Additional logs:However after I closed the error message dialog, the update continue and seems finalized |
|
@chinyongcy Could you please try and paste the resolut of: @Joulinar |
|
short question: should I run the commands before or after updating to 6.26.0?? |
Thank you for your help! Before I went to sleep just now, pass my ssh access to my friend.
I have no idea what it meant, but it fixed the issues. Afterwards, I go to dietpi-software, under software optimised, kodi is no registered as install. EDIT: I tried a clean install, it will still give me the same error as my first post. Have to go through this 3 lines to fix it. |
|
@chinyongcy Another question since you ran "kodi" from console. Does "start_kodi" or "kodi --standalone" work as well? @Joulinar |
|
@MichaIng Typing "kodi --standalone" or "kodi" manage to run the kodi successfully but, it will give me However if I type "start_kodi" anyway what is the difference between "kodi --standalone" and "kodi"? Also another update is, I'm trying to install Netflix on Kodi. I am able to see will this be an issue? and I uninstall kodi from the dietpi-software and just tried your command, I am not sure if it will still helps
After which "journalctl -r" |
|
@chinyongcy
Good to know about fbset still required for this binary, or at least it creates this warnings. You can solve with To replicate the configuration/install error (re-create the files you removed), you need to run For the |
|
@MichaIng I have a theory, not sure if it make sense... Previously when I was not in beta version, I saw your thread on #3031. Follow it last week and manage to install with no issues. Yesterday was fiddling with something and somehow corrupted my SD card. So did a fresh install anyway. Even the solution posted in #3031 will result in the same problem faced. So I am guessing the problem lies with kodi newer not with dietpi? Currently is at version "kodi_18.4-5_buster_all". The previous version is "kodi_18.4-4_buster_all". Not sure if the "4-5" is causing an issue. I tried installing older version by myself, I think I didn't do it correctly. After that I get back Can't seem to get it to work as well. |
|
@chinyongcy However upgrading from an older to a newer package should never cause issues, as this would break every |
|
@MichaIng Yup, everything is working fine, just that the error message is there only. Thank you once again! |
|
@MichaIng Bad news, I'm not able to reproduce the error. Yesterday I had some issues with my HyperV VM and created a new one today, leading to the fact, that all Software title needed to be reinstall from scratch. Installing Wireguard, already forced Anyway, I now have another issue which is not related to the update. After the initial installation of Wireguard, the |
|
@Joulinar
But from your error it is about the PDiffs:
This is hard to replicate since the required PDiffs (differential package lists) depend on the state of the currently loaded lists.
The last two do not contain this I leave now two machines (Buster and Stretch) running until the end of the day and see if I can replicate the issue when some new PDiff is created. In case we need to check whether this is known and expected or otherwise report to Debian bug tracker. So to lower the chance that users run into this during v6.26 update and give us some time to find a solution, I will now move the Sid => Bullseye switch for WireGuard to pre-patches, as well as disabling the package state downloads. Both is already done during regular incremental patches but those run after G_AGUP, thus the "fix" comes after the error terminates the updater. |
|
@MichaIng Yes it seems it depends on how the systems looks before running the update. Unfortunately, I lost my VM yesterday and I'm not able to reproduce the error as well. |
|
@Joulinar I have moved some APT-affecting patches to pre-patches now, so they are applied before G_AGUP and G_AGUG are called. Since "sid" is replaced with "bullseye" for WireGuard, this resolves the issue for the initial call, however it might reappear on a future apt-get update/G_AGUP, so we need to keep eyes open. EDITAnother PDiff has been created, So at least it is not a general incompatibility. Another thing we can do is offering a full APT list reset on G_AGUP failures:
|
|
@MichaIng ahh no no. It was my fault. I deleted my save points on my VM, instead of restoring it. Therefore I lost my v6.25.3 backups. To get back to v6.25.3, it was needed to create new VM from scratch, unfortunately. Means no chance to test on the previous scenario, where it failed. Thx for looking into it :) |
Beta v6.26.1 has been merged: #3141 |
Beta v6.26.2 has been merged: #3145 |
Beta v6.26.3 has been merged: #3155 |
|
Working quite fine with 6.26.3. However, one of my RPi 4s won’t boot with overclock set to 1900mhz and overvoltage 4. It booted without problems with 6.25.3. I think the new boot setting where it boots on max mhz for 20 seconds is the culprit. The same machine boots fine when I set the overclock to 1750 MHz and overvoltage 2. I suppose this specific machine wasn’t as rock stable as I previously thought. |
|
@Vansinnet Could you play around a bid (after doing a backup!), e.g.:
I hope that stress test and |
|
@MichaIng I disabled However, I really like the On my other RPi 4, there's no problem booting with over_voltage 4 @ 1900 mhz even with In the meantime, I think I'll just disable |
|
@MichaIng Conclusion: There's something going on during early boot which makes the RPi sensitive to high overclocks ( Is there anyone else that can verify my finding? |
|
@Vansinnet I just reported the issue to RPi devs: raspberrypi/firmware#1005 |
|
I have seen it reported. |
v6.26 has been released: #3162Many thanks to all testers, I will mark this issue as closed now. |
Important testing cases:
Closes issues:
#1171 #2332 #2641 #2935 #2959 #2962 #2965 #2966 #2975 #2977 #2994 #2996 #2997 #2999 #3007 #3010 #3017 #3024 #3028 #3029 #3031 #3036 #3038 #3045 #3053 #3059 #3060 #3062 #3066 #3070 #3075 #3079 #3085 #3094 #3097 #3098 #3104 #3105 #3111 #3116 #3122 #3123 #3126 #3127 #3128 #3144 #3147 #3151
Known issues
DietPi functionality
SBC/device related
Software title related
The text was updated successfully, but these errors were encountered: