Hi. On Thu, 18 Sep 2014 17:56:00 +0200 Bzzzz <[email protected]> wrote:
> You'd better, as it is even worse: junkD journal is in binary format > and this journal isn't ACID compliant, so any error will render it > unreadable; guess what? It has been rejected as "not a bug" > (https://bugs.freedesktop.org/show_bug.cgi?id=64116); note that > coredumps are also written to this kind of "journal"… Out of curiosity - how exactly being ACID-compliant helps to read a corrupted RDBMS table? What's exactly that bad to keep all coredumps (they're blobs anyway) in one place (a big blob, but it's *one* blob) vs spreading them all other filesystem (s)? I'm not that big fan of binary logs myself, but your arguments seem out of place somehow. Better arguments would be: 1) Unstable journald format. Good luck finding that exact version of journalctl to read logs over next several years. 2) Which leads us to - an upgrade can render all your logs unreadable. That's a real joy, isn't it? 3) Braindead pull-based journald logs replication, as opposed to a conventional push-based one that syslog is using. Reco -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected] Archive: https://lists.debian.org/[email protected]
