Thanks, Andrew. All of our databases do have lc_collate=C. To check my understanding, if that were not the case, we’d need to reindex, but since it is all lc_collate=C, we shouldn’t run into that particular problem. Did I get that right? On Jan 8, 2019, at 3:06 AM, Andrew Gierth <andrew@xxxxxxxxxxxxxxxxxxxx> wrote: >>>>>> "Natalie" == Natalie Wenz <nataliewenz@xxxxxxxxx> writes: > > Natalie> We are in an unusual circumstance, where we need to move all > Natalie> of our Postgres 10 databases from their homes on servers > Natalie> running FreeBSD 11 with ZFS to servers in another data center > Natalie> running Red Hat Linux 7.6 (also with ZFS), with minimal > Natalie> downtime. > > [...] > > Natalie> We’ve tested this procedure on a small database, and it > Natalie> starts, it runs, returns queries. There seems to be no issues. > > Natalie> Is this a reasonable approach? Are there any known > Natalie> concerns/things to investigate before choosing this course? > > Locales will bite you unless your databases all have lc_collate=C or > you're using ICU for all collations); you will need, at minimum, to > reindex every index containing a collatable column with an OS-provided > collation (i.e. not ICU and not C / POSIX). > > -- > Andrew (irc:RhodiumToad)
