Jump to content

WE NEED !YOUR! HELP

Tido

By Tido
in Advanced users - Development

 Share

Recommended Posts

Tido

Tido

Posted
Posted

Hi,

 

Thank you for your interest, you won't be disappointed. You have probably read that the developer want to do some changes - and some of this changes are now ready to get tested in the BETA's of armbian.

 

Do not test with your 'production' SDcard unless you have done a backup ;-)

 

Once done that, please go to 'armbian-config' System  switch to BETA (Nightly), do an update/upgrade, reboot, check if your system still performs as before.

  • If it does work like before - please let us know and post your 'armbianmonitor -u' link
  • If it doesn't work, please try a reboot, check if the problem is reproduceable and if so  explain the steps to reproduce it and post your 'armbianmonitor -u' link

 

If you like to know what has changed or is changing:

Kernel upgrade for Allwinner boards. NEXT is 4.17.y and it's getting daily updates. (LIB_TAG="sunxi-4.18" in our build system)

https://forum.armbian.com/topic/7398-bsp-scripts-rfc/

https://forum.armbian.com/topic/5565-zram-vs-swap/

https://forum.armbian.com/topic/6444-varlog-file-fills-up-to-100-using-pihole/

 

This is not all, see details on GitHub look at the shell code - which makes this happen - more eyes less errors :thumbup:

 

And if you want to try a fresh install, look for the nightly created 17. June https://dl.armbian.com/

 

Thank you in advance for your support

 

2
Tido

Tido

Posted
  • Author
Posted

Hi, I just ran some tests with fresh install on my tinker board

 

  Reveal hidden contents

 

The keyboard-setting seem to stick now and I have installed via armbian-config Transmission and the minimal desktop. After that the Memory footprint in htop went from 70MB to 186MB. File Manager Thunar starts in this release much quicker than with Xenial. The display resolution still perfect on 4.14.

root@tinkerboard:~# armbianmonitor -u
System diagnosis information will now be uploaded to http://ix.io/1dIr

 

Well done :)

 

mpmc

mpmc

Posted
Posted

Not sure what I'm looking for, I've not even noticed any difference. I'm sure there is, but I can't see it. And this is a good thing!

 

I originally did it on my Orange Pi Zero & didn't notice ANYTHING different so did it on my PC and couldn't notice it there either!

 

Update: OK I've found the change to zram on the PC but not the Zero, even after an update!

 

Orange Pi Zero -> http://ix.io/1dIO [Updated] http://ix.io/1dJ3

Orange Pi PC -> http://ix.io/1dIA

 

/me has no idea what to look for.

 

I hope this helps though!

 

 

Igor

Igor

Posted
Posted
  On 6/17/2018 at 10:41 PM, mpmc said:

 

Update: OK I've found the change to zram on the PC but not the Zero, even after an update!

Expand  


This bug will hopefully be fixed with today update ... which I am still fixing for other problems. It would be best to try with not updated image yet or some older image updated to nightly.


Thank you for your help! 

 

Edit: Please pay attention to log rotation service. If that works properly. Starting with today update, current is broken.

1
Tido

Tido

Posted
  • Author
Posted

Shall I stress more the update path to test?

 

  On 6/18/2018 at 4:32 AM, Igor said:

Starting with today update

Expand  

Shall I add in the first post to do: sudo apt update && upgrade for fresh-install to get your latest bugfix?

Tido

Tido

Posted
  • Author
Posted

I just ran an update on :  Armbian_5.38_Tinkerboard_Ubuntu_xenial_next_4.14.xx_desktop.img   'armbian-config' System  switch to Nightly (BETA).

 

My keyboard was now and then in USA-mode instead of Switzerland, so I have started armbian-config. I think this armbian-config (after the upgrade) is different to my fresh install of last night.

It starts now faster just like the fresh install (by the way)  I went to PERSONAL, LOCALES, de_CH.UTF-8  &  en_US.UTF-8, wanted to stay on english (it is easier to copy text to the forum) BUT change to my keyboard - the keyboard change option  did not show up. While I am sure it did on the fresh install - or am I wrong.. it was late.

 

Igor

Igor

Posted
Posted (edited)
  On 6/18/2018 at 7:26 AM, Tido said:

Shall I add in the first post to do: sudo apt update && upgrade for fresh-install to get your latest bugfix?

Expand  

 

Today's update will include more fixes ... but even with it, some leftovers shell be manually cleaned from the previous update. I didn't make a script for that.

 

To do minor, manual clean up:

sudo nano /etc/logrotate.conf

sudo nano /etc/logrotate.d/*

 

Removing those lines:

 

# compress

su root syslog

 

Edited by Tido
changed, took me a bit to understand
Tido

Tido

Posted
  • Author
Posted (edited)
  On 6/18/2018 at 7:54 AM, Tido said:

It starts now faster just like the fresh install (by the way)  I went to PERSONAL, LOCALES, de_CH.UTF-8  &  en_US.UTF-8, wanted to stay on english (it is easier to copy text to the forum) BUT change to my keyboard - the keyboard change option  did not show up. While I am sure it did on the fresh install - or am I wrong.. it was late.

Expand  

just ran  sudo apt update && upgrade   and an update for armbian-config  came through

 

UPDATE:  in the new armbian-config the keyboard is available.

Maybe a stupid question, why didn't it come when I switched to Nightly ?

 

Edited by Tido
keyboard available
Tido

Tido

Posted
  • Author
Posted

 

  On 6/18/2018 at 7:57 AM, Igor said:

/etc/logrotate.d/*

Expand  

my output shows only this files, what is to do ?

 

/etc/logrotate.d/
apt                       armhwinfo                 rsyslog
aptitude                  cups-daemon               unattended-upgrades
armbian-hardware-monitor  dpkg

 

  On 6/18/2018 at 7:57 AM, Igor said:

# compress

su root syslog

Expand  

Wouldn't it be enough to 'comment' these lines, well #compress already is - like this ?

# compress

#su root syslog

 

Igor

Igor

Posted
Posted

Today's updates are made. I only had to purge zram-config to make our zram logging running ... 

apt purge zram-config

 

Tido

Tido

Posted
  • Author
Posted
  On 6/18/2018 at 6:18 PM, Igor said:

I only had to purge zram-config to make our zram logging running

Expand  

that said, does this mean this is obsolete:

  On 6/18/2018 at 7:57 AM, Igor said:

To do minor, manual clean up:

sudo nano /etc/logrotate.conf

sudo nano /etc/logrotate.d/*

 

Removing those lines:

 

# compress

su root syslog

Expand  

 

Tido

Tido

Posted
  • Author
Posted (edited)

I have done the same checks for zram like yesterday. It is one dev less and lzo versus lz4.

 

  Reveal hidden contents

 

hmmm, since I did: apt purge zram-config   reboot   lz4 and the dev came back :-)

  Reveal hidden contents

 

What else can be tested, reported for example log2RAM ?

 

Edited by Tido
forgot - apt purge zram-config
lampra

lampra

Posted
Posted

Espressobin
 

  Reveal hidden contents

 

Igor

Igor

Posted
Posted
  On 6/18/2018 at 11:22 PM, lampra said:

apt purge zram-config

Expand  


That's ok. A package is not present in Debian.

 

Next thing to test is ... logrotate. It check's every 15 minutes if we are over 75%. Then it save logs, rotate them and truncate/zero logs on the /var/log ... this is theory and I hope reality as well. We need to test random, common and extreme situations of logging to see if default settings are O.K. for most cases. Then somewhere must be written how to configure this system.

Tido

Tido

Posted
  • Author
Posted
  On 6/19/2018 at 1:34 PM, Igor said:

ZRAM service timeouts

Expand  

Can you be more specific, how you found that bug?
What can be done to analyse this behavior ?
Is it on both upgrade and fresh install ?

Are the Google results any helpful or is it armbian specific ?

 

Igor

Igor

Posted
Posted
  On 6/20/2018 at 6:55 AM, Tido said:

Can you be more specific, how you found that bug?
What can be done to analyse this behavior ?

Expand  


I was testing freshly made Odroid C2 DEV build with Bionic ... I haven't narrowed down the problem. It can be a kernel or some other service related. ATM I don't know if it's our or general problem.

TonyMac32

TonyMac32

Posted
Posted
  On 6/19/2018 at 1:34 PM, Igor said:

Found troubles on Bionic builds, ZRAM service timeouts.

Expand  

I found the same on Alwwiner H5-based board with Bionic:

  

[FAILED] Failed to start Armbian ZRAM config.
See 'systemctl status armbian-zram-config.service' for details.

That results in a 2 minute boot delay.  It seems to succeed in the end though, or eventually:

  

tony@tritium:~$ sudo systemctl status armbian-zram-config.service
● armbian-zram-config.service - Armbian ZRAM config
   Loaded: loaded (/lib/systemd/system/armbian-zram-config.service; enabled; ven
   Active: active (exited) since Sun 2018-06-24 18:53:02 UTC; 6min ago
  Process: 576 ExecStart=/usr/lib/armbian/armbian-zram-config start (code=exited
 Main PID: 576 (code=exited, status=0/SUCCESS)

Jun 24 18:53:02 tritium armbian-zram-config[576]: no label, UUID=d4c99662-9d44-4
Jun 24 18:53:02 tritium armbian-zram-config[576]: Setting up swapspace version 1
Jun 24 18:53:02 tritium armbian-zram-config[576]: no label, UUID=5b252a9a-1cdf-4
Jun 24 18:53:02 tritium armbian-zram-config[576]: mke2fs 1.44.1 (24-Mar-2018)
Jun 24 18:53:02 tritium armbian-zram-config[576]: [74B blob data]
Jun 24 18:53:02 tritium armbian-zram-config[576]: Creating filesystem with 12800
Jun 24 18:53:02 tritium armbian-zram-config[576]: [41B blob data]
Jun 24 18:53:02 tritium armbian-zram-config[576]: [38B blob data]
Jun 24 18:53:02 tritium armbian-zram-config[576]: [75B blob data]
Jun 24 18:53:02 tritium systemd[1]: Started Armbian ZRAM config.

http://ix.io/1eFg

 

Igor

Igor

Posted
Posted
  On 6/24/2018 at 7:02 PM, TonyMac32 said:

That results in a 2 minute boot delay.  It seems to succeed in the end though, or eventually:

Expand  

 

Yes, it succeeds, but why, for what is waiting? Dependency issues? 

guidol

guidol

Posted
Posted
  On 6/20/2018 at 7:59 AM, Igor said:


I was testing freshly made Odroid C2 DEV build with Bionic ... I haven't narrowed down the problem. It can be a kernel or some other service related. ATM I don't know if it's our or general problem.

Expand  

I did boot Armbian_5.48.180624_Odroidc2_Ubuntu_bionic_dev_4.17.2 today...

I do get a heartbeat at the blue LED, but no sign of life at the ethernet port (green is going up and down, but orange never comes on) ....and I wonder even with a USB-Ethernet device there is no DHCP request.

At this time I didnt got a serial port attached :(

guidol

guidol

Posted
Posted
  On 6/25/2018 at 8:28 AM, guidol said:

I did boot Armbian_5.48.180624_Odroidc2_Ubuntu_bionic_dev_4.17.2 today...

I do get a heartbeat at the blue LED, but no sign of life at the ethernet port (green is going up and down, but orange never comes on) ....and I wonder even with a USB-Ethernet device there is no DHCP request.

Expand  

 

@Igor today I got the same effect (no usb-ethernet) at the OPi Zero Plus2 H5 with 

Armbian_5.48.180624_Orangepizeroplus2-h5_Ubuntu_bionic_dev_4.17.2

 

On my User-Build Kernel (with LED-Support and HDMI-disable for proper reboot) the USB-Ethernet does work.

 

But at the nightly I do get the following error-messages (for orientation some good messages before the errors) via the serial port:
  

Loading Environment from EXT4... ** File not found /boot/boot.env **
** Unable to read "/boot/boot.env" from mmc0:1 **
Failed (-5)
Loading Environment from FAT... Unable to use mmc 1:1... Failed (-5)

[  OK  ] Started Create list of required sta…vice nodes for the current kernel.
[  OK  ] Started Remount Root and Kernel File Systems.
[FAILED] Failed to start Load Kernel Modules.
See 'systemctl status systemd-modules-load.service' for details.
...
[  OK  ] Started Entropy daemon using the HAVEGE algorithm.
         Starting Update UTMP about System Boot/Shutdown...
         Starting Load Kernel Modules...
         Starting Network Time Synchronization...
[FAILED] Failed to start Load Kernel Modules.
See 'systemctl status systemd-modules-load.service' for details.


[    5.041850] brcmfmac: brcmf_fw_alloc_request: using brcm/brcmfmac43430-sdio for chip BCM43430/1
[    5.169329] brcmfmac: brcmf_fw_alloc_request: using brcm/brcmfmac43430-sdio for chip BCM43430/1
[    5.169439] brcmfmac mmc1:0001:1: Direct firmware load for brcm/brcmfmac43430-sdio.clm_blob failed with error -2

Maybe these are the "same" as on the Odroid C2, because there is also no onboard/USB-Ethernet.
Here as hidden the serial Boot-Log from the first boot (my normal user-build system/kernel is on emmc and this is the sdcard boot):
 

  Reveal hidden contents

 

Igor

Igor

Posted
Posted
  On 6/26/2018 at 9:07 AM, guidol said:

the USB-Ethernet does work.

Expand  


As expected. 4.17.y don't have overlay support there ... and USB ports are disabled by default on this board. DEV is experimental/preparation build and many things might be broken.

 

  On 6/26/2018 at 9:07 AM, guidol said:

Maybe these are the "same" as on the Odroid C2

Expand  

 

Unfortunately no. USB on C2 is much more problematic. It's enabled but sometimes doesn't work. Here at least we know why it doesn't :)

 

We seek ZRAM related issues at this point. That is planned to go into stable builds - when it's working as expected or at least if we know what prevents that, while 4.17-18 is months away.

guidol

guidol

Posted
Posted
  On 6/26/2018 at 9:35 AM, Igor said:

As expected. 4.17.y don't have overlay support there ... and USB ports are disabled by default on this board.
DEV is experimental/preparation build and many things might be broken.

Expand  

Ahh - I didnt know that overlay support is disabled - that do clarify some things ;)

I know DEV is a experimental/preparation build, but on the most SBCs here it do work quite well! :)
Now I know that mostly more the onboard/original hardware should work.

 

Igor

Igor

Posted
Posted
  On 6/26/2018 at 10:23 AM, guidol said:

Now I know that mostly more the onboard/original hardware should work.

Expand  


Our target for Allwinner is to get 4.18.y operational and 4.17.y servers only to prepare patches, as a temporal version. 

2
Arglebargle

Arglebargle

Posted
Posted (edited)

Can someone answer a naive question about the new zram implementation? I think I've found something important that's been overlooked. 

  

# Use half of the real memory by default --> 1/${ram_divisor}
	ram_divisor=2
	mem_info=$(LC_ALL=C free -w 2>/dev/null | grep "^Mem" || LC_ALL=C free | grep "^Mem")
	memory_total=$(awk '{printf("%d",$2*1024)}' <<<${mem_info})
	mem_per_zram_device=$(( ${memory_total} / ${zram_devices} / ${ram_divisor} ))

	for (( i=1; i<=zram_devices; i++ )); do
		[[ -f /sys/block/zram${i}/comp_algorithm ]] && echo lz4 >/sys/block/zram${i}/comp_algorithm 2>/dev/null
		echo -n ${mem_per_zram_device} > /sys/block/zram${i}/disksize
		mkswap /dev/zram${i}
		swapon -p 5 /dev/zram${i}
	done
	echo -e "\n### Activated ${zram_devices} zram swap devices with ${mem_per_zram_device} MB each\n" >>${Log}

 

Is the intent here to devote half of real memory to zram and take advantage of compression to multiply that chunk of memory, or is the intent to provide some small swap space without hitting swap on flash storage?

 

Right now the script is doing the latter, not the former. The current implementation makes swap space equal in size to half of ram, which is then compressed (usually at 2:1 or 3:1, most often at 3:1) and stored in memory. This means that, on average, the amount of actual ram used by zram hovers between 25% and ~17% rather than the apparently intended 50%.

 

I spent quite a bit of time researching and profiling something similar for my own use and I've settled on the following to allocate 50% of real memory to zram:

 

    mem_per_zram_device=$(( 3 * ${memory_total} / ${zram_devices} / ${ram_divisor} ))

 

Coincidentally this is the same zram allocation that Google makes on ChromeOS.

 

I've seen the same mistake made in almost every single zram script that I can find via Google so this definitely isn't an Armbian thing. It looks like someone made the original logic error at some point a long time ago and it's been copy-pasted into every zram setup script for the last 7 or 8 years.

 

For what it's worth setting /sys/block/zramN/mem_limit does what everyone seems to be expecting and limits the amount of ram that the zram device can use. Unfortunately there is no way to adaptively vary /sys/block/zramN/disksize according to the actual compression ratio so the currently accepted best practice seems to be assuming 3:1 and punting (this is what Google does on ChromeOS.)

Edited by Arglebargle
clarification
Flávio Becker

Flávio Becker

Posted
Posted

Hi, I just installed the 5.49 nightly build (4.17.3) on my OrangePI Prime.

It's running quite well and I had almost no issues until now, except for:

. some errors reported by journald (journal corruption) that ended up filling up all /var/log space.  

. sometimes the reboot doesn't work, freezing on a black screen with a cursor blinking in one corner.

The errors reported by journald were probably caused by my need to power off and on to resume after the reboot freezing.

I had to delete the log files though to be able to perform the system settings (language, timezone and so on). 

 

Igor

Igor

Posted
Posted
  On 6/29/2018 at 5:40 PM, Flávio Becker said:

The errors reported by journald were probably caused by my need to power off and on to resume after the reboot freezing.

Expand  


Not sure. I also noticed this (only?) on Bionic version which has some other problems and it's still in testing/fixing phase.

 

Thank you for the report. Next time supply logs with: 

armbianmonitor -u

This way we can see more.

Guest
This topic is now closed to further replies.
 Share
Go to topic listing
×
×
  • Create New...

Important Information

Terms of Use - Privacy Policy - Guidelines

  I accept