Hi all, > Am 30.01.2025 um 12:35 schrieb A FreeBSD User <free...@walstatt-de.de>: > Well, in this case I do not use dedicated ZIL drives. I also made several > experiences with > "single" ZIL drive setups, but a dedicated ZIL is mostly useful in cases were > you have > graveyard full of inertia-suffering, mass-spinning HDDs - if I'm right the > concept of SSD > based ZIL would be of no use/effect in that case. So I ommited tose.
Dedidated SLOG vdevs only make sense if your workload involves synchronous writes like when running databases or VM virtual disks over NFS or iSCSI. > Every single SSD originally plugged in is now back in place, even the faulted > one (which > doesn't report any faults at the moment). > > Although the pool isn't "importable", zdb reports its existence, amongst > zroot (which resides > on a dedicated drive). There a various things yet to try, but I have no personal experience with them: - "force" read-only import - import referencing a past transaction group, i.e. a previous version of the entire pool How to find these transaction groups - no idea, sorry. >>> The host is running the lates Xigmanas BETA, which is effectively >>> FreeBSD 14.1-p2, just for the record. This topic including sometimes successful recovery frequently pops up on the TrueNAS forum. Did you ask in the Xigmanas community? Weren't they able to help? I know there are regulars on the TrueNAS forum who know what to do. Maybe head over there? As I wrote there are threads going like: "please provide the output of `zdb something someting`" *output* "now to `zpool import something something`" ... etc. HTH, Patrick
