log2ram

ryecoaaron

Is it recommend to do this if running a helio4 from a omv4to5 upgrade:

# disable armbian log services if found

for service in log2ram armbian-ramlog armbian-zram-config; do

if systemctl list-units --full -all | grep ${service}; then

systemctl stop ${service}

systemctl disable ${service}

fi

done

rm -f /etc/cron.daily/armbian-ram-logging

if [ -f "/etc/default/armbian-ramlog" ]; then

sed -i "s/ENABLED=.*/ENABLED=false/g" /etc/default/armbian-ramlog

fi

if [ -f "/etc/default/armbian-zram-config" ]; then

sed -i "s/ENABLED=.*/ENABLED=false/g" /etc/default/armbian-zram-config

fi

if [ -f "/etc/systemd/system/logrotate.service" ]; then

rm -f /etc/systemd/system/logrotate.service

systemctl daemon-reload

fi

I did this manually and onced my /var/log folder2ram sized increased from 100MB to 1000MB.

Thanks

My logs are not getting that big i noticed the folder sized was sized to 100MB then went to 1000MB

folder2ram 1008M 32M 977M 4% /var/log

My question how will logs get rotated so it will not fill up?

Thanks

Zitat von PEW

My question how will logs get rotated so it will not fill up?

Don't you have logrotate running?

No.. Because the install script will remove it based because this a helios4. I did and upgrade from omv4 to 5 so i was wondering how logs are rotated.

i first noticed logrotate was not working then came upon this thread and read the install script which i believe is in the armbian-config utility, But, I did an upgrade not a fresh install based on this thread:

https://forum.armbian.com/topi…-helios4-support/page/22/

I am now trying to figure out random reboots i am having.

https://forum.armbian.com/topi…-helios4-support/page/23/

I just disabled the watchdog service. I then listed all the services and noticed the logrotate service not working correctly then found what the install script does.

So now i wondering how logs will rotate so it will not fill up.

# disable armbian log services if found

for service in log2ram armbian-ramlog armbian-zram-config; do

if systemctl list-units --full -all | grep ${service}; then

systemctl stop ${service}

systemctl disable ${service}

fi

done

rm -f /etc/cron.daily/armbian-ram-logging

if [ -f "/etc/default/armbian-ramlog" ]; then

sed -i "s/ENABLED=.*/ENABLED=false/g" /etc/default/armbian-ramlog

fi

if [ -f "/etc/default/armbian-zram-config" ]; then

sed -i "s/ENABLED=.*/ENABLED=false/g" /etc/default/armbian-zram-config

fi

if [ -f "/etc/systemd/system/logrotate.service" ]; then

rm -f /etc/systemd/system/logrotate.service

systemctl daemon-reload

fi

Thanks

Yes. I was just wondering if this unit file (logrotate systemd) is removed if it will still do log rotation. I did notice the logrotate cron job in cron.daily but I noticed this in the logrotate script:

if [ -d /run/systemd/system ]; then

exit 0

fi

i have this directory so i am wondering if it will just exit and never do any log rotation.

Thanks

i copied the logrotate.service file back for systemctl and logrotate is now working.

I was getting this error before i went down this path of disabling /dev/zram and converting to folder2ram:

error: /etc/logrotate.conf:18 duplicate log entry for /var/log/wtmp

error: /etc/logrotate.conf:25 duplicate log entry for /var/log/btmp

This was caused by have wtmp and btmp in the /etc/logrotate.d folder and in the logrotate.conf so i commented out the btmp and wtmp statements in the logrotate.conf

I am just trying to fix issues i have been noticing since this latest release:

cron jobs failing after 5.5.8-1 update - armbian-check-first-login.sh

Thanks

Thanks for the reply. Yes, I understand.

These were things i just noticed. my omv-backup cron job stopped working and the cron job wasn't finding the correct path /usr/sbin. I am justing to pin point what happened.

It was around the Aug 28 timeframe.

One thing i noticed is that if i create a cron job for omv-backup, it will not run unless i give it a certain time not @daily.

0 0 * * * root /var/lib/openmediavault/cron.d/userdefined-137e4193-ba57-4674-a99a-24f55b8e9a7e >/dev/null 2>&1

Do you like the helios64?

I think i am going to wait for the next batch with ddr memory.

Thanks

Hey ryecoaaron,

I'm having the same question as DimaKompot.

Would you mind leaving an answer?

I'm getting OOM errors on a regular basis on my NanoPi M4 and try to figure out whether I should enable normal swap or change

/etc/default/armbian-zram-config to be enabled.

Zitat von DimaKompot

Hello ryecoaaron,

I'm using Ordoid XU4 (Raspbian). I installed a new fresh system today with OMV on top of it with your installation script. And now I don't have any /dev/zram device and also don't have any swap anymore. Went through latest changes and find that topic. Is it an expected behavior?

Code

root@odroidxu4:~# lsblk
NAME         MAJ:MIN RM  SIZE RO TYPE MOUNTPOINT
sda            8:0    0  5.5T  0 disk 
mmcblk0      179:0    0 29.1G  0 disk 
└─mmcblk0p1  179:1    0 28.8G  0 part /
mmcblk0boot0 179:16   0    4M  1 disk 
mmcblk0boot1 179:32   0    4M  1 disk 

Code

root@odroidxu4:~# free
              total        used        free      shared  buff/cache   available
Mem:        2043820      191824     1533760       28384      318236     1761480
Swap:             0           0           0

Thank you for your reply!

My server is usually idling at 700 MB usage out of 2 GB RAM.

OOMs occur every 2-5 days at 2 a.m. and I guess it's PLEX doing maintenance jobs.

I'm totally happy using swap as the slowdown should only occur at night where it doesn't bother me.

Would you suggest to use armbian-specific zram implementation in /etc/default/armbian-zram-config which also enables swap from my understanding, or the generic linux approach?

I'll try the generic then. Appreciate your input