Hi Oswald, On Wed, Oct 02, 2024 at 11:33:54AM GMT, Oswald Buddenhagen wrote: > the most plausible explanation is that you have a concurrent background > sync job. maybe triggered by a directory watcher on the mailboxes, which > would cause the race somewhat reliably.
Ahhh, you're right. On one of the computers I added a cron job recently. I had forgotten. However, the other one doesn't have a cron job. Maybe I just run it from one terminal and forgot about it and run it from another terminal... If I see that happen again on the system that doesn't have a cron job, I'll let you know. > > isync uses process-associated locking, so an internal origin is > impossible even you are trying to sync the same channel/box multiple > times and isync somehow forgets to close the state files. > > isync doesn't check the errno of the failed fcntl call, so it's possible > that the error message is bogus. strace would tell. > > are you running an experimental kernel? maybe it's being funny ... I'm on Sid. It should be good, but you never know. :) Cheers, Alex -- <https://www.alejandro-colomar.es/>
signature.asc
Description: PGP signature
