It's probably fixed by https://github.com/dovecot/core/commit/3001c1410e72440e48c285890dd31625b7e12555
> On 25. Nov 2024, at 14.51, Bjørnar Ness via dovecot <dovecot@dovecot.org> > wrote: > > Here are some numbers from a test run: > > dst: full pull sync from src > 6m19s > dst: delta pull sync from src > 2s > > > src: full push dync to destination > 4h14m34s > src: delta push sync to destination > 6s > > > So the initial sync is 4h14m (for push) vs 6m20s (for pull) > > Can someone answer what is causing this very odd behavior? > > > man. 18. nov. 2024 kl. 13:18 skrev Bjørnar Ness <bjornar.n...@gmail.com>: >> >> I have been experimenting with dsync lately, and to my surprise I found that >> dsync in forward direction is extremely slow compared for reverse (-R) >> direction for the exact same data, disk caches flushed both times. I have >> created a docker-compose setup to prove my findings. Am I missing something >> here, or is this a bug? >> >> link to tests: https://github.com/bjne/dovecot-dsync-benchmark >> >> -- >> Bj(/)rnar > > > > -- > Bj(/)rnar > _______________________________________________ > dovecot mailing list -- dovecot@dovecot.org > To unsubscribe send an email to dovecot-le...@dovecot.org _______________________________________________ dovecot mailing list -- dovecot@dovecot.org To unsubscribe send an email to dovecot-le...@dovecot.org
