2) is TRIM planned? Quite important thing i think. No TRIM roughly equals
of running SSD with all filesystems full.
Yes, someday, though we ran into a few issues trying to make it work
the way we wanted.
That's what i wanted to hear. I am not in a hurry. All i do now on FreeBSD
works just fine and i do not change working thing in any case.
But i already understood that in not so long future FreeBSD will not be
usable while older versions would not work on new hardware or will have
too many unfixed bugs or some new needed software will not compile from
ports on it.
If TRIM will work a year from now it is already great.
Thank you very much for answer!
My recommendation would be to only partition some
of the disk, leaving additional spare area. Somewhat ironically, I
Right i already know this and it seems fine as TEMPORARY solution.
think ssd controllers have gotten better at garbage collection on
their own, so trim support is perhaps slightly less important than it
No matter how well it works, if they know what is unused it never gets
copied to new blocks.
So TRIM may not improve speed but always improve wear.
Imagine one filesystem designed for flash media, sort of BSD LFS-alike.
you put one single file on it, run vnconfig and then create UFS on it and mount.
TRIM is ability to "punch holes" in that single file. More unused
space=less block rewrites when cleaner works.
That's how today SSD works, with that "vnconfig" part handled by "SSD
controller". It is fast, but it actually mean how fast it
really could be without double filesystem layer, if we could just buy
packaged pile of FLASH chips and one interface chip doing just ECC, with
flash dedicated filesystem running directly on OS.