Good day. I have have an offline pool that appears to have missing devices. I seem to have narrowed it down to the write cache disks and it appears that both my disks have failed. i have replaced both the disks but the pool is still offline. i assume that I need to somehow add it to the configuration and replace the old write cache disks but cannot seem to get it done or find the options for it.
Welcome to TrueNAS and it’s forums!
First, ZFS does not have write cache disks. If you meant SLOG / LOG, then yes, failure of both would prevent normal importation of the pool.
It would be best to copy and paste the output here of zpool import
to be sure that is the problem.
The command option to import a pool that has failed SLOG device(s) is below. Note that this can cause most recent synchronous writes to be lost, permanently. However, if your pool was gracefully exported, (or server gracefully rebooted), this should not happen.
zpool import -m POOL_NAME
Now some details. ZFS requires a pool to be imported to replace any devices. Just randomly replacing a device, like SLOG, when the pool is exported, does nothing to fix the problem.
Next, failure of a SLOG during normal operation is fine. It will impact synchronous write usage, since ZFS we then use ZIL, (data vDev intent log), instead of the now failed Separate intent LOG. But no impact to data or data integrity.
Further, SLOG devices only help for specific cases, synchronous writes, like for NFS or iSCSI. Outside of that, SLOG devices are not really needed or used. Your use of “write cache” indicates that you may not have known this detail.
Last, SLOG devices generally want or need some special characteristics;
- High endurance
- Low capacity is all that is needed. A 64GB or larger SSD / NVMe is wasted, unless that is all you have that meets the other wants and requirements
- High write speed
- Power loss protection
Must read: