Short answer: not too much
and the few is different
On my way from (Open)Solaris, its free forks to Open-ZFS on BSD, then Linux, I have accepted that every step reduces the easyness and stability of a Sun Solaris with one ZFS version on one OS, managed with two commands, with OS/ZFS kernelintegrated iSCSI, NFS and SMB to a bunch of distributions, each with a different Linux, different ZFS release or SAMBA and different setup or update methods or bug states and the worst are special ZFS APIs + databases instead the simple and always up to date zfs and zpool commands.
And now Qnap
When I tried to check out if a Qnap ZFS NAS can be a managed member of a napp-it cs ZFS serverfarm among other ZFS servers on Linux, Unix or Windows, I did not expect problems as I thought Qnap is simply another Linux + Open-ZFS + Web-gui appliance. It is not.
Setup of napp-it_cs on Qnap was easy. After installing Perl from the Qnap repo it worked but only to show some ZFS stats. ZFS manipulation was not possible, nor disks shown as Qnap lacks lsblk to show disks.With the help of p. who has a spare Qnap i tried to find out which Open-ZFS version is used (for compatibility). A zpool version only shows 5000 instead for exemple Open-ZFS 2.2.x as expected, It seems Qnap wants to hide the ZFS release.
A zpool get all shows that it must be a quite old Open-ZFS without any of the newer ZFS features like for example zstd, allocation_classes, checkpoints, device_removal, raidz_expansion or draid, needed for compatibiliity (pool move, replication) with other Open-ZFS systems
My next thougt was, this cannot be as I remembered that Qnap proudly announced Raid-Z Expansion a year ago, long befor it became stable in Open-ZFS 2.2.3. I also remembers that Qnap calls the Raid on ZFS Raid 5/6. I thought first this was an error of someone translating the docs but now I checked again and it seems indeed that Qnap is using ZFS ontop Raid 5/6 with a huge amount of unique global and private ZFS properties. A Pool move from/to Qnap is propably impossible and you loose auto repair due checksum errors and safe atomic writes or in the end just like ZFS ontop hardware raid.
On Qnap docs I found: QNAP Flexible Storage Architecture | QuTS hero 5.0.x
This is in no way a ZFS layout, more like a modified mdadm setup with some extras. It is clear now, why Qnap called its feature raidz_expand while the Open-ZFS feature has the name raidz_expansion. It is not the same feature but a transformation for a Raid 5/6. This repeats with many other features that are Qnap only while all newer fancy Open-ZFS features are missing.
I will no disqualify. Synology did the same with btrfs ontop Linux mdadm raid that therefor lacks modern filesystm features like self healing filesystems or safe atomic writes due Copy on Write with btrfs/ZFS Raid instead oldstyle Raid 1/5/6. Such helps for the easy switch of an older ext4 GUI concept to btrfs or ZFS.
Remains the question: Who wants this half baked solutions without newer core ZFS advantages.
ZFS is printed ontop the box in huge letters and you may expect the real ZFS is inside.
The exact amount of differences to a Open-ZFS philosophy and many internal technical details are not published, but ZFS?
Maybe in name only. It may be quite good, easy to manage and quite safe (but not as safe as regular ZFS configs)
some stats, compare with yours!
and the few is different
On my way from (Open)Solaris, its free forks to Open-ZFS on BSD, then Linux, I have accepted that every step reduces the easyness and stability of a Sun Solaris with one ZFS version on one OS, managed with two commands, with OS/ZFS kernelintegrated iSCSI, NFS and SMB to a bunch of distributions, each with a different Linux, different ZFS release or SAMBA and different setup or update methods or bug states and the worst are special ZFS APIs + databases instead the simple and always up to date zfs and zpool commands.
And now Qnap
When I tried to check out if a Qnap ZFS NAS can be a managed member of a napp-it cs ZFS serverfarm among other ZFS servers on Linux, Unix or Windows, I did not expect problems as I thought Qnap is simply another Linux + Open-ZFS + Web-gui appliance. It is not.
Setup of napp-it_cs on Qnap was easy. After installing Perl from the Qnap repo it worked but only to show some ZFS stats. ZFS manipulation was not possible, nor disks shown as Qnap lacks lsblk to show disks.With the help of p. who has a spare Qnap i tried to find out which Open-ZFS version is used (for compatibility). A zpool version only shows 5000 instead for exemple Open-ZFS 2.2.x as expected, It seems Qnap wants to hide the ZFS release.
A zpool get all shows that it must be a quite old Open-ZFS without any of the newer ZFS features like for example zstd, allocation_classes, checkpoints, device_removal, raidz_expansion or draid, needed for compatibiliity (pool move, replication) with other Open-ZFS systems
My next thougt was, this cannot be as I remembered that Qnap proudly announced Raid-Z Expansion a year ago, long befor it became stable in Open-ZFS 2.2.3. I also remembers that Qnap calls the Raid on ZFS Raid 5/6. I thought first this was an error of someone translating the docs but now I checked again and it seems indeed that Qnap is using ZFS ontop Raid 5/6 with a huge amount of unique global and private ZFS properties. A Pool move from/to Qnap is propably impossible and you loose auto repair due checksum errors and safe atomic writes or in the end just like ZFS ontop hardware raid.
On Qnap docs I found: QNAP Flexible Storage Architecture | QuTS hero 5.0.x
This is in no way a ZFS layout, more like a modified mdadm setup with some extras. It is clear now, why Qnap called its feature raidz_expand while the Open-ZFS feature has the name raidz_expansion. It is not the same feature but a transformation for a Raid 5/6. This repeats with many other features that are Qnap only while all newer fancy Open-ZFS features are missing.
I will no disqualify. Synology did the same with btrfs ontop Linux mdadm raid that therefor lacks modern filesystm features like self healing filesystems or safe atomic writes due Copy on Write with btrfs/ZFS Raid instead oldstyle Raid 1/5/6. Such helps for the easy switch of an older ext4 GUI concept to btrfs or ZFS.
Remains the question: Who wants this half baked solutions without newer core ZFS advantages.
ZFS is printed ontop the box in huge letters and you may expect the real ZFS is inside.
The exact amount of differences to a Open-ZFS philosophy and many internal technical details are not published, but ZFS?
Maybe in name only. It may be quite good, easy to manage and quite safe (but not as safe as regular ZFS configs)
some stats, compare with yours!
Code:
scan: none requested
prune: never
expand: none requested
config:
NAME STATE READ WRITE CKSUM
zpool1 ONLINE 0 0 0
raidz1-0 ONLINE 0 0 0
qzfs/enc_0/disk_0x1_5000C500E5D7D6AC_3 ONLINE 0 0 0
qzfs/enc_0/disk_0x2_5000C500E5D7D701_3 ONLINE 0 0 0
qzfs/enc_0/disk_0x3_5000C500E5CB2C85_3 ONLINE 0 0 0
errors: No known data errors
[admin@NAS66BC3A tmp]# zpool get all
NAME PROPERTY VALUE SOURCE
zpool1 size 54.4T -
zpool1 qsize 38892364682836 -
zpool1 orig_qsize 39196211281920 -
zpool1 capacity 0% -
zpool1 altroot - default
zpool1 health ONLINE -
zpool1 guid 1560259431169208497 default
zpool1 qguid LN3VG3CX default
zpool1 version - default
zpool1 bootfs - default
zpool1 delegation on default
zpool1 autoreplace off default
zpool1 cachefile - default
zpool1 failmode continue default
zpool1 listsnapshots on local
zpool1 autoexpand on local
zpool1 globalcache off default
zpool1 globalcache_notuser off default
zpool1 l2rebuild off default
zpool1 dedupditto 0 default
zpool1 dedupratio 1.02x -
zpool1 free 54.4T -
zpool1 allocated 318M -
zpool1 dspace 35.4T -
zpool1 dedup-saving 1204224 -
zpool1 max_poolop 60 -
zpool1 readonly off -
zpool1 comment - default
zpool1 expandsize 0 -
zpool1 freeing 0 default
zpool1 qthresh 80% local
zpool1 qthreshsize 31006227903283 local
zpool1 overqthresh no local
zpool1 qthreshavail 7751556975820 local
zpool1 qsnap 7751556975820 local
zpool1 usedbysnapshot 37654336 -
zpool1 owner 0 default
zpool1 upsecs 0 default
zpool1 upsecsupdate 0 default
zpool1 throttle on default
zpool1 raidz_layout layout_reorder default
zpool1 raidzshift 24 default
zpool1 qos_enable 0 default
zpool1 qos_max_4kbase 0 default
zpool1 qos_weight 0 default
zpool1 qos_throttle off default
zpool1 qos_reserved 0 default
zpool1 smartddt_loadratio 0 default
zpool1 smartddt_txgdirty 0 default
zpool1 smartddt_txgdirty_low 0 default
zpool1 smartddt_times 0 default
zpool1 smartddt_entrydrops 0 default
zpool1 resilver_ratio 50 local
zpool1 scrub_ratio 50 local
zpool1 smartddt_state 0 -
zpool1 smartddt on default
zpool1 l2cache_ioalign off default
zpool1 indirectlayout on local
zpool1 shadowminshift 12 default
zpool1 shadowashift 12 default
zpool1 shadowblockshift 18 default
zpool1 zib_size 34730608794010 -
zpool1 orig_zib_size 38606387281920 -
zpool1 zib_free 34730553711002 -
zpool1 zib_allocated 55083008 -
zpool1 zib_metasize 1511828488192 -
zpool1 zib_metafree 1511809802240 -
zpool1 zib_worstamp 2 -
zpool1 pool_overprovision 10 local
zpool1 compdedup_count 0 -
zpool1 compdedup_maxcount 8388608 default
zpool1 compdedup_minpshift 9 default
zpool1 zib_falloc_size 1048576 -
zpool1 zib_falloc_txg 4 -
zpool1 ssd_overprovision 0 default
zpool1 qlog_policy legacy default
zpool1 prune_goal 10000000 default
zpool1 ddt_entry_limit 20000000 default
zpool1 prune_slack_txg 32 default
zpool1 async_write_min_active 0 default
zpool1 logvolume - default
zpool1 async_write_max_active 16 local
zpool1 async_read_min_active 0 default
zpool1 async_read_max_active 0 default
zpool1 sync_write_min_active 0 default
zpool1 sync_write_max_active 0 default
zpool1 sync_read_min_active 0 default
zpool1 sync_read_max_active 0 default
zpool1 prune_goal_by_ram on default
zpool1 prune_goal_deduced 18500000 -
zpool1 deadlist metadspace 272800 default
zpool1 deadlist datadspace 0 default
zpool1 raidzshift_i 24 local
zpool1 aggrprefetch off default
zpool1 aggrprefetch_maxinit_sz 67108864 default
zpool1 nomal_class_metadspace 1006901395456 default
zpool1 nomal_class_datadspace 39818373365760 default
zpool1 ddt_freq_on_disk off -
zpool1 ddt_prune_percentage 10 -
zpool1 ddt_dec_freq_sec 1209600 -
zpool1 ddt_prune_min_time_ms 50 -
zpool1 ssd_life_type off default
zpool1 shadow_refmap_shift 36 default
zpool1 shadow_refmap_reserve_shift 28 default
zpool1 ssdop_size 0 -
zpool1 reserved_size 151397597184 -
zpool1 asynccow on local
zpool1 resilver_pause off default
zpool1 scan_ignore_error off default
zpool1 tag QNAPxxxxxxxxx local
zpool1 qsal_retention_data 0 -
zpool1 qsal_retention_freed 0 -
zpool1 qsal_retention_spill 0 -
zpool1 vdev_aggregation_limit 0 default
zpool1 zib_disable_ref_negative_verify 0 default
zpool1 migrate_error_handle continue default
zpool1 migrate_ratio 0 default
zpool1 spacelow_thresh 8% default
zpool1 spacelow_threshsize 3100622790328 local
zpool1 spacelow_overthresh no local
zpool1 spacelow_threshavail 17179869184 local
zpool1 spacelow_threshbyte 17179869184 default
zpool1 feature@async_destroy enabled local
zpool1 feature@empty_bpobj active local
zpool1 feature@lz4_compress active local
zpool1 feature@encryption enabled local
zpool1 feature@raidz_layout enabled local
zpool1 feature@raidz_shift enabled local
zpool1 feature@extensible_dataset enabled local
zpool1 feature@meg_blocksize enabled local
zpool1 feature@sha512 enabled local
zpool1 feature@skein enabled local
zpool1 feature@edonr enabled local
zpool1 feature@indirect_layout active local
zpool1 feature@zib_async_destroy enabled local
zpool1 feature@clog enabled local
zpool1 feature@ddt_prune enabled local
zpool1 feature@raidz_shift_i active local
zpool1 feature@deadlistv2 active local
zpool1 feature@zibddt_prune enabled local
zpool1 feature@ssd_lifeop active local
zpool1 feature@indirect_layout_bp_fill active local
zpool1 feature@asynccow active local
zpool1 feature@large_dir enabled local
zpool1 feature@raidz_expand enabled local
zfs get all zpool1/zfs1
NAME PROPERTY VALUE SOURCE
zpool1/zfs1 type filesystem -
zpool1/zfs1 creation Wed Jul 31 22:17 2024 -
zpool1/zfs1 used 1.03G -
zpool1/zfs1 available 989M -
zpool1/zfs1 referenced 35.2M -
zpool1/zfs1 compressratio 1.81x -
zpool1/zfs1 mounted yes -
zpool1/zfs1 quota none default
zpool1/zfs1 reservation none default
zpool1/zfs1 recordsize 128K local
zpool1/zfs1 mountpoint /share/ZFS1_DATA local
zpool1/zfs1 sharenfs off default
zpool1/zfs1 checksum on default
zpool1/zfs1 compression on local
zpool1/zfs1 atime rel default
zpool1/zfs1 devices on default
zpool1/zfs1 exec on default
zpool1/zfs1 setuid on default
zpool1/zfs1 readonly off default
zpool1/zfs1 zoned off default
zpool1/zfs1 snapdir visible local
zpool1/zfs1 aclmode passthrough local
zpool1/zfs1 aclinherit passthrough local
zpool1/zfs1 canmount on local
zpool1/zfs1 xattr on default
zpool1/zfs1 copies 1 default
zpool1/zfs1 version 5 -
zpool1/zfs1 utf8only off -
zpool1/zfs1 normalization none -
zpool1/zfs1 casesensitivity mixed -
zpool1/zfs1 vscan off default
zpool1/zfs1 nbmand off default
zpool1/zfs1 sharesmb off default
zpool1/zfs1 refquota 1G local
zpool1/zfs1 refreservation 1G local
zpool1/zfs1 primarycache all default
zpool1/zfs1 secondarycache none local
zpool1/zfs1 usedbysnapshots 219K -
zpool1/zfs1 usedbydataset 35.2M -
zpool1/zfs1 usedbychildren 0 -
zpool1/zfs1 usedbyrefreservation 1024M -
zpool1/zfs1 logbias latency default
zpool1/zfs1 dedup off local
zpool1/zfs1 mlslabel -
zpool1/zfs1 sync standard local
zpool1/zfs1 refcompressratio 1.81x -
zpool1/zfs1 written 82.7K -
zpool1/zfs1 encryption off -
zpool1/zfs1 keysource none default
zpool1/zfs1 keystatus none local
zpool1/zfs1 physicalused 34.7M -
zpool1/zfs1 physicalreferenced 34.6M -
zpool1/zfs1 logicalused 63.1M -
zpool1/zfs1 logicalreferenced 62.9M -
zpool1/zfs1 txg 0 -
zpool1/zfs1 qthresh 50% local
zpool1/zfs1 qthreshsize 536870912 -
zpool1/zfs1 overqthresh no -
zpool1/zfs1 wormtype off default
zpool1/zfs1 wormtrigger admin default
zpool1/zfs1 wormretention 0y0m1d default
zpool1/zfs1 defaultuserquota none default
zpool1/zfs1 defaultgroupquota none default
zpool1/zfs1 datatrace off default
zpool1/zfs1 wormguid 8602461902382197696 -
zpool1/zfs1 wormdiff 0 default
zpool1/zfs1 snapcount 2 local
zpool1/zfs1 redundant_metadata most default
zpool1/zfs1 fastcopy on local
zpool1/zfs1 dedup_activated off -
zpool1/zfs1 cachewritten auto default
zpool1/zfs1 qos_enable 0 default
zpool1/zfs1 qos_max 0 default
zpool1/zfs1 qos_min 0 default
zpool1/zfs1 qos_weight 0 default
zpool1/zfs1 qos_priority 0 default
zpool1/zfs1 qos_burst 0 default
zpool1/zfs1 qos_burst_time 0 default
zpool1/zfs1 qos_allocated 0 -
zpool1/zfs1 wormprivilege off default
zpool1/zfs1 wormlog no default
zpool1/zfs1 acltype richacl default
zpool1/zfs1 wormcommit none default
zpool1/zfs1 aggrprefetch on default
zpool1/zfs1 archives_counting off default
zpool1/zfs1 thin_shared_byte 32K -
zpool1/zfs1 compact_byte 0 -
zpool1/zfs1 snap_refreservation 0 default
zpool1/zfs1 usedbysnaprsrv 0 -
zpool1/zfs1 overwrite_reservation 35.2M -
zpool1/zfs1 large_dir on local
zpool1/zfs1 large_dir_threshold 9.77K default
zpool1/zfs1 clonescount 0 local
zpool1/zfs1 qnap:zfs_volume_name System local
zpool1/zfs1 qnap:zfs_flag 257 local
zpool1/zfs1 qnap:zfs_threshold 50 local
zpool1/zfs1 qnap:zfs_guid 1560259431169208497-1722457073-zfs1 local
zpool1/zfs1 qnap:zfs_edge_cache_type 0 local
zpool1/zfs1 qnap:share_path System local
zpool1/zfs1 qnap:snapshot_reserved_bytes 0 inherited from zpool1
zpool1/zfs1 qnap:pool_flag 1 inherited from zpool1
zpool1/zfs1 qnap:snapshot_reserved 20 inherited from zpool1
zpool1/zfs1 qnap:snapshot_reserved_enable 2 inherited from zpool1
Last edited:
