On Wed, 16 Aug 2006 05:33:26 -0500 bingo <[email protected]> wrote:
> It would be good if logtail supports locking. I think we need some more information if this bug is to be action-ed. logcheck uses logtail2 now (and syslog is not the default):so perhaps it is not relevant after nearly 20 years (there were other replies requesting patches in the intervening period, and no-one has provided them -- so perhaps we should just close this as no longer relevant). If the issue is that rsyslog might still be writing to the file, when logtail runs, surely all that might happen is that the same entries might be re-checked on the next run? (and I dont think it would be sensible for logtail to try and stop rsyslog writing while logtail runs) the journal has a better way to avoid races, using --cursor-file -- logcheck doesnt yet use that (watch this space!) which i think might also avoid potential race conditions here
