Posts by ginsa

    Code
    systemctl status logrotate.timer
    ● logrotate.timer - Daily rotation of log files
    Loaded: loaded (/lib/systemd/system/logrotate.timer; enabled; vendor preset: enabled)
    Active: active (waiting) since Sun 2020-12-27 09:57:54 CET; 3h 2min ago
    Trigger: Mon 2020-12-28 00:00:00 CET; 10h left
    Docs: man:logrotate(8)
    man:logrotate.conf(5)
    Dec 27 09:57:54 arxiv systemd[1]: Started Daily rotation of log files.

    Maybe not...

    Code
    systemctl status logrotate.service
    ● logrotate.service - Rotate log files
    Loaded: loaded (/etc/systemd/system/logrotate.service; static; vendor preset:
    Active: inactive (dead)
    Docs: man:logrotate(8)
    man:logrotate.conf(5)

    After the command systemctl start logrotate:

    What can i do ?

    Thank you for your advice.

    I used the installation script. The plugin folder2ram is running.

    I always shut downed the system using the function of "Scheduled jobs" turning it off at 00.00 and turn it on at 09.00 in the next morning:

    0 0 * * * root rtcwake -m off -s 32400

    it should guarantee a correct shutdown with syncing the syslog on flash drive.

    I have not zram

    Code
    lsblk
    NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT
    sda 8:0 0 1.8T 0 disk
    ├─sda1 8:1 0 10G 0 part
    └─sda2 8:2 0 1.8T 0 part /srv/dev-disk-by-label-DATA
    mmcblk1 179:0 0 29.8G 0 disk
    ├─mmcblk1p1 179:1 0 16G 0 part /
    └─mmcblk1p2 179:2 0 13.5G 0 part

    The strange thing is that for 2 months (September - November) I can retrieve the 2000 pages of system log that suddenly it disappeared from 10/11/2020 at boot...

    After that date I only upgraded the system but I did not change manually any configuration.

    As I already said I can see the current syslog with journalctl but not by the webUI. "System logs" function seems not to work anymore.

    As this seems to be an OS issue, sharing infos about deatils before you installed OMV would increase likelihood of replies

    OMV 5.5.19-1 now runs on Odroid HC2 with Armbian 20.11.3 Buster (Linux 4.14.212-odroidxu4).

    I installed the system on September with Ambian 20.08 that was subsequently updated.

    I followed the standard procedure and scripts to install the Armbian and OVM 5 without any problem and warnings.


    A useful info might be if other users with the same configuration had that kind of problems.

    What files or configuration can I check ?

    On my system I cannot see the System logs the web function seems not to work correctly.

    It claims there are 2001 pages, the first on 18th September, the last is on 10th November. I can't see more recent pages.

    I noticed that the syslog was growing without any rotate so I tried to reinstall the system, but after the result was the same.

    With journalctl command I can see the today log, but If I try to see the yesterday boot log the system replies "Specifying boot ID or boot offset has no effect, no persistent journal was found."

    The system resides on a SD card... But why are the October/November pages present and not yesterday's?

    On directory /var/log there are 2 files:

    -rw-r----- 1 root adm 18688229 Dec 26 11:45 syslog

    -rw-r----- 1 root adm 1004948 Sep 18 00:00 syslog.1


    syslog file seems to be big enough to contain more than the daily records.


    Is this the correct behavior ? Before the last date (10/11/2020) I could see the log of previous days in the System log function.


    Thanks for any help.

    I tried to get by on my own trying to understand a little more:

    So logrotate doesn't work because the scripts and the configuration files are corrupted and this was what I suspected.

    But if I try to uninstall logrotate, to reinstall it again hoping it rebuilts those files, I get the warning:

    "The following packages will be REMOVED:

    logrotate openmediavault openmediavault-flashmemory openmediavault-omvextrasorg openmediavault-resetperms openmediavault-usbbackup"

    So logrotate is strictly connected to OMV modules.

    This makes me think who may have changed those scripts and configuration files ....


    I think that the only solution to my problem is to start over by installing Armbian, OMV and Nextcloud.

    I am certainly not very skilled, but a software that constantly forces me to reinstall the whole system is not the best !!

    Probably thousands of users will disagree because they are completely satisfied, but after more than 2 months of "beta testing" I still haven't been able to put the services into "production". I just keep reinstalling everything all over again, until to meet the next problem...

    Hi, OMV WebGUI today says that my syslog is 680 pages, 34000 records, and the last and only time he did logrotate was 9 days ago. I reinstalled the system recently.

    All the files in /etc/logrotate.d have the date of today boot.

    if I have a look into the files many of them (but not all) seem corrupt, here are some:

    It seems there are strange characters (00 in hexadecimal) inside the files that truncate the scripts. Some file starting with "{" don't have the ending "}".

    I don't know why, because the syslog management is completely in charge to Armbian that define the base rules.

    Is there a way to bypass this strange situation and allow to rotate the logs correctly, once a week and storing them in a compressed way ?

    I hope it is not necessary to rebuilt the system from scratch, I have done this too many times already.

    I reinstalled OMV 5.5.9-1on Armbian 20.08 and it seem the scheduled jobs don't seem to work anymore.

    Here attached the image of the GUI of the command set to switch off the system at midnight and switch it on at 9.00 in the morning.

    The mail sent at execution has subject:

    [localhost] Cron <root@arxiv> /var/lib/openmediavault/cron.d/userdefined-d419c9d7-e6e4-4f75-bdd0-cb8f5ab0f848 | mail -E -s "Cron - Power off @00.00 Power on @9.00" -a "From: Cron Daemon <root>" root >/dev/null 2>&1

    The command output in the mail message is:

    /var/lib/openmediavault/cron.d/userdefined-d419c9d7-e6e4-4f75-bdd0-cb8f5ab0f848: 20: /etc/profile.d/armbian-check-first-login.sh: Syntax error: "(" unexpected

    It reports an error executing a system script (armbian-check-first-login.sh). The file /var/lib/openmediavault/cron.d/userdefined-....0f848 however contains the correct command "rtcwake -m off -s 32400".

    If I try to run in OMV the scheduled job I always get a message error but the rtcwake command correctly works.

    Is here any mistake ? How can I resolve it ?

    I don't think changing Docker Storage has any effect on already installed dockers. I don't think it moves anything. I assume it has to be changed before you do anything at all with docker, to work as intended.

    It works in general, but once Portainer fails to start, the other time Nextcloud doesn't start, sometimes both correctly work ... the system is no longer reliable as before.

    Conclusion: it is better to avoid these changes on the fly and to put Docker on the right place before to install the containers .... :thumbdown:;)

    Check to be sure there is no docker subdirectory under /var/lib. Or at least that it isn't used.


    I don't think changing Docker Storage has any effect on already installed dockers. I don't think it moves anything. I assume it has to be changed before you do anything at all with docker, to work as intended.

    I copied SD directory /var/lib/docker to a new directory on the disk with rsync.

    I renamed the old library as /var/lib/docker.old, so it disappeared from the active system.

    I changed Docker Storage on GUI and rebooted OMV.

    Portainer didn't start, but Nextcloud regularly worked.

    I saw that changing Docker Storage it changes accordingly the content of /etc/docker/daemon.json

    I think now Docker no longer uses anything on the SD.

    Great, it works. Nextcloud works without reinstalling and that is very important.

    I had problems with Portainer, but but it took few seconds to reinstall it.

    The new docker directory must have the same attributes as the old.

    Wonderful ! Now is too late but tomorrow I'll surely do it. Thank you very much.

    That wasn't a good recipe. Now Docker and then Nextcloud and Portainer don't work anymore ... ;(

    It seemed too simple. Luckily I have a copy of the system on cloned SD card and I can restart immediately. 8o

    Waiting for tips...

    Found on the net...


    Stop Docker service

    Edit the file /etc/docker/daemon.json changing the string "data-root": "/var/lib/docker" to "data-root": "/path/to/your/new/docker"

    Copy the Docker library: rsync -aP /var/lib/docker/ /path/to/your/new/docker

    Rename the old directory /var/lib/docker

    Restart Docker service

    If everything works, remove the old directory /var/lib/docker, otherwise restore all changes made


    It seems easy and safe, I'll try.

    Thank you for your advices.

    ......

    If you HDD will die, your data as well as your system will die too. So you definitely need to reinstall completely.

    Unfortunately it happened... so I decided to keep the system on SD card that I can periodically clone.

    ...........

    Create a separate share or shares on the HDD for docker images and docker app data.

    This way the SD card will see very little writes, protected by the flash memory plugin, and can be expected to last many years.

    The docker images and docker app data are likely to see a lot of writes, and also expand over time. Not suitable to store on the SD card!

    .......

    I think you refer to use of the command "docker run parameters /directory/on/hard/disk appname". I did it so.

    Because with the command "docker pull appname" I don't know if I can define where the image have to be installed in a location different from /var/lib/docker on the SD card.

    I am using OMV5 on Odroid HC2 with 2 TB disk, I also installed Nextcloud.

    My simply question: is it better to maintain the system on SD card as now or create a partition of ten of GB on the disk and move with the utility the system on it ?

    I have no problem of performance, my only goal is the services reliability and continuity.

    In this way it is easy to clone the system on another card, but is the system on SD with the flashmemory plugin reliable enough over time.

    How did others who have a configuration similar to mine choose?

    Thanks.