FreeBSD 10.1 Released

[zfrogz]

Good news! It has now matured to 10.1 which means it might be stable enough for the development team to start looking at it.
https://www.freebsd.org/releases/10.1R/announce.html

Here is the stuff this community is probably most interested in:

The vfs.zfs.zio.use_uma sysctl(8) has been re-enabled. On multi-CPU machines with enough RAM, this can easily double zfs(8) performance or reduce CPU usage in half. It was originally disabled due to memory and KVA exhaustion problem reports, which should be resolved due to several changes in the VM subsystem. [r260338]

A new flag, -R, has been added to the fsck_ffs(8) utility. When used, fsck_ffs(8) will restart itself when too many critical errors have been detected. [r260178] (Contributed / provided by Netflix)

The zfs(8) filesystem has been updated to implement “bookmarks”. See zfs(8) for further details. [r263407]

The zfs(8) filesystem has been updated to allow tuning the minimum “ashift” value when creating new top-level virtual devices (vdevs). To set the minimum ashift value, for example when creating a zpool(8) on “Advanced Format” drives, set the vfs.zfs.min_auto_ashift sysctl(8) accordingly. [r266122]

The libzfs thread pool API has been imported from OpenSolaris, and adapted for FreeBSD. This change allows parallel disk scanning, which can reduce zpool(8) overall import time in some workloads. [r266612]

The restore(8) utility has been updated to prevent assertion failures when restoring a UFS filesystem dump to a ZFS filesystem by writing restored files in block sizes that are a multiple of 1024. [r269651]

Two sysctl(8)s have been added to the zfs(8) filesystem: [r269774]

• vfs.zfs.mg_fragmentation_threshold: The percentage of the metaslab group size that should be considered eligible for allocation, unless all metaslab groups within the metaslab class have also crossed this threshold.
  
  
• vfs.zfs.metaslab.fragmentation_threshold: The maximum percentage of metaslab fragmentation level to keep their active state
  

The default zfs(8) ARC hash table size has been increased, and a new loader(8) tunable, vfs.zfs.arc_average_blocksize, has been added. Previously, the hash table could be too small, which would lead to long hash chains and limit performance for cached reads. The vfs.zfs.arc_average_blocksize tunable allows overriding the default block size. The previous default was 65536, and default of the new loader(8) tunable is 8192. [r269846]

 

[bestboy]

The vfs.zfs.zio.use_uma sysctl(8) has been re-enabled. On multi-CPU machines with enough RAM, this can easily double zfs(8) performance or reduce CPU usage in half. It was originally disabled due to memory and KVA exhaustion problem reports, which should be resolved due to several changes in the VM subsystem. [r260338]

Code:

[root@Cygnus] /# sysctl vfs.zfs.zio
vfs.zfs.zio.use_uma: 1

Seem's to be enabled by default on 9.2.1.8. So either it was not disabled on FreeBSD 9 or it has been reenabled for FreeNAS?

 

[cyberjock]

Code:

[root@Cygnus] /# sysctl vfs.zfs.zio
vfs.zfs.zio.use_uma: 1

Seem's to be enabled by default on 9.2.1.8. So either it was not disabled on FreeBSD 9 or it has been reenabled for FreeNAS?

Because FreeNAS is an appliance and not an OS, you can't take things that are said about FreeBSD and blindly expect them to be true for FreeNAS. The FreeNAS devs have already made specific optimizations to tunables, sysctls, and to the code itself that make FreeNAS more of a one-of-a-kind OS. In fact, FreeNAS 9.3 has bookmark support. ;)