Something seems broken with iSCSI since 6.7.0

Please send in fresh logs.

I have the same problem.

My ReadyNas is solely used as iSCSI Target.

I have created a LUN with 3.269TB (90% of Space) and activated Bit-Rot Protection for this LUN.

root@NAS:~# btrfs fi du /data/RB00_READYNAS02/.iscsi/
     Total   Exclusive  Set shared  Filename
   3.27TiB     3.19TiB           -  /data/RB00_READYNAS02/.iscsi//iscsi_lun_backing_store
     0.00B       0.00B           -  /data/RB00_READYNAS02/.iscsi//.serial_number
   3.27TiB     3.19TiB    75.14GiB  /data/RB00_READYNAS02/.iscsi/

root@NAS:~# btrfs fi df /data
Data, single: total=3.62TiB, used=3.62TiB
System, DUP: total=8.00MiB, used=480.00KiB
Metadata, DUP: total=5.00GiB, used=4.16GiB
GlobalReserve, single: total=512.00MiB, used=0.00B

root@NAS:~# du -h --max-depth=2 /data | sort -hr
3.3T    /data/RB00_READYNAS02/.iscsi
3.3T    /data/RB00_READYNAS02
3.3T    /data
1.7M    /data/.apps
1.6M    /data/.apps/.freeapps
76K     /data/.apps/.readydlna
20K     /data/._share
16K     /data/._share/RB00_READYNAS02
0       /data/.vault
0       /data/.timemachine
0       /data/.purge
0       /data/home
0       /data/.apps/.xdg
0       /data/.apps/.forked-daapd
0       /data/.apps/DO_NOT_DELETE

Where have the missing 350GiB gone?

Best Regards

Got it!

The iSCSI File is heavily fragmented (over 1 Million Extents!)

I disabled CoW for the file (chattr +C <file>) and defragmented extents smaller than 100M (btrfs fi defrag -f -t 100M /data/RB00_READYNAS01/.iscsi/iscsi_lun_backing_store).

When enough space was freed up (>50GB) I stopped defragmentation and started again but this time for extents smaller than 10G.

I hope the file won't fragment again this heavy after i disabled CoW.

You can't disable CoW on existing files that contain data.  So the chattr command you ran had no effect.

You may be able to improve fragmentation characteristics by adding the "autodefrag" mount option to the /etc/fstab entry.

Well, well, well.

iscsi_lun_backing_store: 1616701 extents found

Even though repeated defragging has freed up some space on the drive (835 GB "free" as of now), the originally reported problem still stands: when "sync writes" are disabled, TrueCrypt iSCSI volumes are no longer mountable since 6.7.0.

The issue is still present with 6.8.0 RC1.

Of course, with sync writes on the performance is also truly horrible and the device is practically unusable.

I am eating the consequences of the many bizarre decisions NTGR have taken during the last few years (essentially destroying the potential of capable hardware with horrible software decisions) and will be copying all data off the device, downgrade to OS4 (or fashion some way of forcing Debian on it), and never again buy or recommend a Netgear product in my life.

This ends a 20 year+ old relationship, back to the time where "Bay Networks" was on Netgear products. Good work lads.