Tom Mueller wrote:
Shawn Walker wrote:
2) that you do not have a pkg.depotd process running that is pointed
at the repository indicated by file://
Is a pkgsend/pkgrecv into a file:// repo ok when a pkg.depotd is running
with --readonly, followed by a SIGUSR1 on the pkg.depotd after the
publish is complete?
Yes, the --readonly scenario should work correctly. The primary concern
is with having a read/write pkg.depotd serving /repo/foo while
publishing using file:///repo/foo at the same time.
In short, only one process should be writing to the repository at the
same time.
If not, what is the use case where the SIGUSR1 is useful?
Ignoring the not part, there are multiple use cases for SIGUSR1:
* an update of the <repo_dir>/cfg_cache file
* the --readonly publication case above
* a filesystem change (a zfs rollback or changing the symlink that was
used for a repository directory perhaps?) that means you want the
repository to see new package, catalog, or configuration data
Cheers,
--
Shawn Walker
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
