I currently have a 2-server setup (3 servers in total, one still to be configured for this) with one server using systemd-journal-remote
(.socket
& .service
) to collect all log entries from the other machine, which uploads with systemd-journal-upload
.
I have installed version 239 via stretch-backports
on the receiving debian server, simply to make journalctl -m
work, and pick up /var/log/journal/remote
This setup is for redundancy and simplicity of logging, and to have a overview of all events on all servers, but systemctl -f -m
does not pick up all logging origins (local and remote), can this be fixed, or is it via design?
Edit:
Upon inspection with a rather non-reoccurring service (cron
), i finally got entries for both machines, but here I think i find the problem;
{ ... "__REALTIME_TIMESTAMP": "1539710221237971", "__MONOTONIC_TIMESTAMP": "5621984625", "_BOOT_ID": "afb0d4143851464184f340c4ace9XXXX", "_MACHINE_ID": "59b666a7337442898dc7cc671c0eXXXX", "_HOSTNAME": "SERVER_RECEIVE", ... }, { ... "__REALTIME_TIMESTAMP": "1539706141105212", "__MONOTONIC_TIMESTAMP": "6030701064304", "_BOOT_ID": "afb0d4143851464184f340c4ace9XXXX", "_MACHINE_ID": "35ade424e59d4ef18a9986a090f6XXXX", "_HOSTNAME": "SERVER_SEND", ... },
journalctl
seems to sort on __MONOTONIC_TIMESTAMP
, is there a reason as to why the remote capturing service seems to iterate on a lower count than the actual journalctl
count?