8 8
Igor

testers wanted Next major upgrade v5.60

Recommended Posts

https://github.com/armbian/documentation/blob/master/docs/Release_Changelog.md (please add what you think I forgot and is relevant to mention, up to 6 months back)

  • all images were rebuilt, except boards which support ended,

and with a few exceptions which are still in the works (Tinkerboard, MiQi and perhaps we can move T4/RockPRO to the stable) and will be added probably next week. I already test a few and so far I couldn't find troubles. I record them here: https://github.com/armbian/testings. The result is displayed in this https://beta.armbian.com/buildlogs/report.html table and when it is filled up, an update to the main repository on all images can be rolled out.

Share this post


Link to post
Share on other sites

i'm not sure if this is relevant, armbian stretch orange pi one

i switch to nightly and dev kernel

umbian-monitor -u

http://ix.io/1mhS

 

symptom:

after the kernel switch from 4.14 (stretch) to 4.18 (dev)

2 things seem to fail

1) avahi-daemon

avahi-daemon is running on the board but when i issue avahi-browse -a from host pc

the response:

+ enp3s0 IPv6 orangepione [02:81:c4:64:0d:8a]               Workstation          local

is missing, on orange pi one avahi-daemon is setup for
publish-workstation=yes
, there is no change in the config prior to the update and after

 

if i restart avahi-daemon on the board:

sudo systemctl restart avahi-daemon.service

things work again

 

2) rpi-monitor

 

rpi-monitor failed to start

 

$ sudo /usr/bin/rpimonitord
Failed to create share at /usr/bin/rpimonitord line 161.

>>

  $this->{'sharedmem'} = IPC::ShareLite->new(
        -key     => $this->{'daemon'}->{'sharedmemkey'}, <<< line 161
        -create  => 'yes',
        -destroy => 'no'
    ) or die $!;

<<

tl;dr not checking further for now, i'd check again probably sometime later

Share this post


Link to post
Share on other sites
6 hours ago, ag123 said:

 

Oh, there's something we also need to take care of with upgrades: the existing 128MB swap file on SD card present in /etc/fstab:

              total        used        free      shared  buff/cache   available
Mem:           494M         66M        269M        2.7M        158M        415M
Swap:          375M          0B        375M

NAME       ALGORITHM DISKSIZE  DATA COMPR TOTAL STREAMS MOUNTPOINT
/dev/zram0                50M 11.1M  2.9M  3.3M       4 /var/log
/dev/zram1 lz4          61.8M    4K   63B    4K       4 [SWAP]
/dev/zram2 lz4          61.8M    4K   63B    4K       4 [SWAP]
/dev/zram3 lz4          61.8M    4K   63B    4K       4 [SWAP]
/dev/zram4 lz4          61.8M    4K   63B    4K       4 [SWAP]

With vm.swappiness=100 now it might be possible that the system decides to start to swap to SD card even if swap priority is pretty low. @Igor what do you think? Checking size of swap file and if it's 'our' 128MB file then removing fstab entry and swap file on upgrade?

Share this post


Link to post
Share on other sites

rebooted after a night:

something is a little unusual with umbian-monitor -u

 

8 hours ago, ag123 said:

 

umbian-monitor -u

http://ix.io/1mhS
 

Sun Sep 9 03:22:31 +08 2018 | Orange Pi One | 5.59.180906 | armhf | armv7l | 4.14.68-sunxi ### dmesg: [ 0.000000] Booting Linux on physical CPU 0x0 [ 0.000000] Linux version 4.14.68-sunxi (root@nightly) (gcc version 7.2.1 20171011 (Linaro GCC 7.2-2017.11)) #161 SMP Thu Sep 6 22:43:53 CEST 2018

 

if i do this on the linux command line:

 

Quote

$ uname -a
Linux orangepione 4.18.6-sunxi #159 SMP Thu Sep 6 22:57:30 CEST 2018 armv7l GNU/Linux
 

$ dmesg | head
[    0.000000] Booting Linux on physical CPU 0x0
[    0.000000] Linux version 4.18.6-sunxi (root@nightly) (gcc version 7.2.1 20171011 (Linaro GCC 7.2-2017.11)) #159 SMP Thu Sep 6 22:57:30 CEST 2018

i repeated sudo ambianmonitor -u

http://ix.io/1mjj

the same symptom, linux version did not match and the time did not seem to match either

^^^^ edit:

the dmesg puzzle is resolved:

kernel version 4.18 is showing near the bottom of armbianmonitor -u log

 

the prior 2 symptoms

1) avahi-daemon

somehow it works this first time round avahi-browse -a from host pc detected it
+ enp3s0 IPv4 orangepione [02:81:c4:64:0d:8a]               Workstation          local
symptom is erratic

2) rpi monitor

$ sudo /usr/bin/rpimonitord
Failed to create share at /usr/bin/rpimonitord line 161.

this failed the same way

 

@tkaiser

 

$ /sbin/sysctl -A |grep vm.swappiness

vm.swappiness = 60

$/sbin/swapon
NAME      TYPE SIZE USED PRIO
/var/swap file 128M   0B   -2

$ /sbin/zramctl

returns empty

 

zram did not instantiate

 

and on that swapfile, i think for orangepione, 512M memory, it's good to leave that swapfile there.
running 'desktop' with a web browser would hit that even with zram enabled. either way the swapfile can be added later by users

 

Share this post


Link to post
Share on other sites
13 minutes ago, ag123 said:

something is a little unusual


Log has a complete history ... from where you started. Scroll down.

 

1 hour ago, tkaiser said:

what do you think? Checking size of swap file and if it's 'our' 128MB file then removing fstab entry and swap file on upgrade?


Sounds O.K., will prepare and test.

Share this post


Link to post
Share on other sites

thanks igor

 

$ grep SHMEM /boot/config-4.18.6-sunxi
CONFIG_SHMEM=y

 

$ sudo /usr/bin/rpimonitord

Failed to create share at /usr/bin/rpimonitord line 161.

>>

  $this->{'sharedmem'} = IPC::ShareLite->new(
        -key     => $this->{'daemon'}->{'sharedmemkey'},  <<< line 161
        -create  => 'yes',
        -destroy => 'no'

<<

 

the rpimonitord failed due to shared memory error would take more time to figure out

the kernel config is apparently configured with shared memory enabled
it seemed to be simply a rpimonitor issue, don't worry about it for now non-critical there is still armbianmonitor -m

- signing off for now

Share this post


Link to post
Share on other sites

hi just a feedback about the rpimonitord issue
, i just did a new install for a new sd card orange pi pc debian stretch (Armbian 5.59)

after i switched to the nighly development builds (Armbian 5.59.180909 nightly)

i've verified that rpimonitor is still running

the kernel is Linux orangepipc 4.14.69-sunxi #183 SMP

but when i switched to the development kernel as well 4.18.7-sunxi #180 SMP

rpimonitor failed

 

then i go into /boot and did

diff -C 2 config-4.14.69-sunxi config-4.18.7-sunxi |less

i observe this

Spoiler

*** 48,63 ****
  CONFIG_DEFAULT_HOSTNAME="(none)"
  CONFIG_SWAP=y
! CONFIG_SYSVIPC=y
! CONFIG_SYSVIPC_SYSCTL=y
! CONFIG_POSIX_MQUEUE=y
! CONFIG_POSIX_MQUEUE_SYSCTL=y
  CONFIG_CROSS_MEMORY_ATTACH=y
! CONFIG_FHANDLE=y
! CONFIG_USELIB=y
! CONFIG_AUDIT=y
  CONFIG_HAVE_ARCH_AUDITSYSCALL=y
- CONFIG_AUDITSYSCALL=y
- CONFIG_AUDIT_WATCH=y
- CONFIG_AUDIT_TREE=y
 
  #
--- 50,59 ----
  CONFIG_DEFAULT_HOSTNAME="(none)"
  CONFIG_SWAP=y
! # CONFIG_SYSVIPC is not set
! # CONFIG_POSIX_MQUEUE is not set
  CONFIG_CROSS_MEMORY_ATTACH=y
! # CONFIG_USELIB is not set
! # CONFIG_AUDIT is not set
  CONFIG_HAVE_ARCH_AUDITSYSCALL=y
 
  #

 

the various parameters for CONFIG_SYSVIPC, CONFIG_POSIX_MQUEUE, CONFIG_USELIB, are apparently switched off in the 4.18 kernel builds orange pi pc (h3)

 

i've verified that rpimonitor functionality is restored by reverting the kernel to 4.14 using armbian-config

Linux orangepipc 4.14.69-sunxi #183 SMP Sun Sep 9 23:38:42 CEST 2018 armv7l GNU/Linux

 

armbianmonitor -u ( just prior to downgrading the kernel to 4.14.69)

http://ix.io/1mp9

 

 

Share this post


Link to post
Share on other sites
1 minute ago, ag123 said:

've verified that rpimonitor functionality is restored reverting the kernel to 4.14 using armbian-config


So you are saying there is a problem in kernel configuration? Well, 4.18.y had some issues with memory handling and I had to disable certain advanced functions (mainly Docker dependencies). Perhaps a bit too much was removed ... we had "relocations out of range" in the logs and some modules failed to load ...

 

Anyway. 4.18.y is as is and doesn't need to be more polished for the next release. Eventually.

Share this post


Link to post
Share on other sites
3 minutes ago, ag123 said:

apparently i think rpimonitor use sysv-ipc (i.e. shmem) hence as it is turned off in the config in 4.18 builds, that probably breaks rpimonitor


Possible. If you can, try to enable those things back and monitor logs for suspicious errors ... which might also not be Allwinner related problem. I didn't get to the bottom of the problem yet.

Share this post


Link to post
Share on other sites

for now i think it is ok as after all i'm running a 'development' kernel, i'd try to setup my build environment as diskspace etc are running low on my host pc. hence i mainly run the images on my board and i'm happy to try out the 'development' images and kernels

i hope providing some feedback could help in evaluating the (kernel) configs etc so that we may 'preempt' some issues for the 'user' release

if you enable the configs CONFIG_SYSVIPC, CONFIG_POSIX_MQUEUE, CONFIG_USELIB, i'd try again to switch to 4.18 kernel in armbian-config

Share this post


Link to post
Share on other sites

Sorry if this is in wrong topic and/or the questions are too stupid. I am quite new to armbians inner workings, but would like to contribute anyway.

I was trying to create test report for OrangePi plus as instructed in https://github.com/armbian/testings. My intention is to create the test reports for OrangePi plus as they seem to be missing and also to test things mentioned in the first post of this topic. I am somewhat confused by the instructions. Is it preferred to run the createreport.sh on all variants of Armbian i.e. Armbian_5.59_Orangepiplus_Debian_stretch_next_4.14.65.img, Armbian_5.59_Orangepiplus_Ubuntu_bionic_next_4.14.65.img and from them by installing the nightly from armbian-config? Totaling 4 test reports, one for each kernel?

How about the tests mentioned in the first post? Only bug's and issues reported here considering the first post, right?

Oh, should the process of changing to nightly image, started from armbian-config sit in the "Configuring base-files" 16% for over an hour? If not, how do I debug it? I am beginning to think it's stuck, as my CPU is idling too. (From armbianmonitor: 12:42:34:  480MHz  0.06   1%   0%   0%   0%   0%   0% 40.3°C  0/9). I am using Armbian_5.59_Orangepiplus_Ubuntu_bionic_next_4.14.65 image.

I am doing this over SSH, as I don't have suitable KB for this. For the armbian-config I am using serial connection trough UART-USB bridge and the armbianmonitor is run from SSH over WLAN.

 

Oh. Regarding the DVFS, is it enough that I can see the CPU frequency changing in armbianmonitor or are there other tests I should do? No, I am not going to measure the core voltages by multi meter. ;)

Edited by mdm63
More questions.

Share this post


Link to post
Share on other sites
8 minutes ago, mdm63 said:

Is it preferred to run the createreport.sh on all variants of Armbian i.e. Armbian_5.59_Orangepiplus_Debian_stretch_next_4.14.65.img, Armbian_5.59_Orangepiplus_Ubuntu_bionic_next_4.14.65.img and from them by installing the nightly from armbian-config? Totaling 4 test reports, one for each kernel?

well not really. for now, let's assume that rootfs stuff works and we only test for new kernels bootloaders. So for each kernel you need one test. :) Current work-flow could be:

  1. Burn Ubuntu/debian default, next or dev image to SD-Card
  2. switch to nightly in armbian-config (so that your board gets updates from beta.armbian.com instead of apt.armbian.com)
  3. git clone https://github.com/armbian/testings
    cd testings
    ./createreport.sh

    e.g. USB can be tested by attaching a USB stick and check dmesg, dvfs can be tested with stress -c4 and running armbianmonitor in a second console

  4. you should be prompted through all stuff you've to answer

In case there're still open questions feel free to ask. 

 

39 minutes ago, mdm63 said:

I am doing this over SSH, as I don't have suitable KB for this. For the armbian-config I am using serial connection trough UART-USB bridge and the armbianmonitor is run from SSH over WLAN.

you can also have multiple ssh sessions to the same board if you need more than one console (e.g. for dvfs tests). When possible please test connection over ethernet cable as well. 

 

I suggest that you do this on an SD-card without productive stuff in case something goes wrong and the update to nightly bricks your board just to avoid stress on your side in case it breaks. 

Share this post


Link to post
Share on other sites

armbian creating armbianEnv.txte after first boot

 

Hi, this is my first post in this forum, I don't know if this is the correct thread, but I just trying fresh Armbian 5.59 next on nanopim1, orangepi zero dan nanopiduo. After first boot, armbianEnv.txt content is replaced with these :

#	$OpenBSD: sshd_config,v 1.101 2017/03/14 07:19:07 djm Exp $

# This is the sshd server system-wide configuration file.  See
# sshd_config(5) for more information.

# This sshd was compiled with PATH=/usr/bin:/bin:/usr/sbin:/sb

and armbianEnv.txte file is created with content from original armbianEnv.txt. I verify that armbianEnv.txte is not exist on sd card just after flashing.

I tried with 4 Sandisk brand and I can replicate the phenomenon. Is this a bug ?

 

 

Share this post


Link to post
Share on other sites

I retried above case again with orange pi zero using armbian bionic and stretch. The phenomenon happen on both image. Additional things that I found out is that :

after first boot : armbianEnv.txte created, armbianEnv.txt still has original content

after second boot: armbianEnv.txt content replaced with sshd_config

Share this post


Link to post
Share on other sites
51 minutes ago, Nuha Arina Rafiuddin said:

Will the image download updated to 5.60 too, or it is supposed to be "apt upgrade" from 5.59 ?

There is no need to rebuild images. Upgrade, yes.

Share this post


Link to post
Share on other sites
1 hour ago, Nuha Arina Rafiuddin said:

Thanks, I am confused because I see release notes for 5.60 but image in dl.armbian.com only have 5.59

The repository has been updated to v5.60

So you have to download V5.59 (or less) and apt update & apt upgrade to v5.60 from the repository.

Maybe the images would be updated at another time or you have to build the image for yourself via the armibian-build-system - if you need a image which has v5.60 from the start.

Share this post


Link to post
Share on other sites

did an

apt-get update

followed by

apt-get upgrade

from 5.59 nightly builds on orange pi pc

upgrades apparently went through successfully despite some errors post the follow-up scripts for zram-config, the errors seemed safe to ignore, rebooted and kernel is upgraded to 4.14.70-sunxi by the upgrade. things seemed to work just well in 4.14.70

 

further switched to development kernel 4.18.8-sunxi via armbian-config

the current set of things seemed to work just well so far, rpimonitor now works in 4.18.8 no issues

currently running via ssh remotely, i've not yet tested desktop

 

thanks @Igor et.al

-------

minor note, found that in 4.18.8 cpu seem to run only between 2 frequencies 648mhz, 1ghz

hence i did a check between the dtb files (decompiling it using dtc -I dtb -O dts /boot/dtb/sun8i-h3-orangepi-pc.dtb)

it turns out 4.18.8 do not have the operating point definitions that is there in 4.14.70

however, the dtb in 4.18.8 has some new definitions for csi (this may be good news for those wanting to try out the camera in the mainline kernel)

various other definitions also varies between 4.14.70 vs 4.18.8

my guess is 4.14.70 could be considered stable and 4.18.8 still considered development/experimental

switched back to 4.14.70

htmi-sound is also missing

 

Share this post


Link to post
Share on other sites

the last months I did use the image for the NanoPi Neo Plus2 on my NanoPi Neo Core2.

Now that there is a "real" image for the Core2 I reinstalled completly - but got a "strange" message about my 32GB EVO MicroSD Card:

"Attention: Your SD card seems to be very slow. Please check performance using armbianmonitor -c" :(

 

A Samsung 32GB EVO slow? but anyway -  I did now install via armbian-config to eMMC and the message is gone :)

Share this post


Link to post
Share on other sites
4 hours ago, guidol said:

A Samsung 32GB EVO slow? but anyway

Yes possible, I've one or two Samsung 32GB EVOcards which also show up a warning. F3 shows no corrupted sectors, seq write for large garbage is also okay but 4k is horrible.. 

Share this post


Link to post
Share on other sites
22 minutes ago, ag123 said:

interestingly an nslookup on ix.io

https://network-tools.com/nslook/Default.asp?domain=ix.io&type=255&server=8.8.8.8&class=1&port=53&timeout=5000&go.x=15&go.y=3

Queried domain does not exist

 

maybe it is temporary

 

Domain Name: IX.IO

Registrar WHOIS Server: whois.101domain.com

Updated Date: 2018-09-20T07:04:58Z

 

DNS servers have been changed today, probably just waiting for propagation although it should have happened by now. If you query against NS1.CHUNKHOST.COM instead of google DNS the domain is valid and active and you will see what the new records are. Strangely despite the change happening several hours ago there has been no propagation to any other major DNS systems

Share this post


Link to post
Share on other sites
7 minutes ago, guidol said:

root@npi-neo-core2:~# armbianmonitor -u
System diagnosis information will now be uploaded to http://ix.io/1n6C

 

The card is here:

### quick iozone test: 4 1175 2216 8181 8187 8813 186

That's just 186/4 --> ~45 random write IOPS with 4k blocksize. Seriously way too low for good rootfs performance. The Samsung EVO/EVO+ recommendation isn't valid since 2017 any more. A1 rated cards are the only things to buy.

 

What's also interesting: 'Armbian ramlog partition with lz4 compression' turned into lz4hc later. I need to check this again. On the other hand once the kernels are upgraded to 4.18 or above we'll have zstd anyway so most probably not worth the efforts to analyze...

 

 

Share this post


Link to post
Share on other sites

ok testing out desktop x-windows typing from armbian ...   nightly upgraded to 4.14.70 kernel
started x-windows light-dm
terminal works, thunar file manager works
chromium (old news) and firefox crashes :lol:
no worries for the hardcore there is still netsurf web browswer ;)
oh audio over hdmi works on orange pi pc, so do opi pc's analog audio ;)
mpv video works with audio over hdmi
editors leafpad,geany works

 

armbianmonitor -u

https://pastebin.com/3NLDHaFs

(truncated due to size limits, the latest entry is at bottom)

 

chromium.log

firefox.log

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
8 8