zfs-0.7.11-8llnl

Changes from the prior release:

• Minor fix to debug patch for TOSS-4454

zfs-0.7.11-7llnl

Changes from the prior release:

• Additional debug patch for the issue with dmu_zfetch (TOSS-4454)

zfs-0.7.11-6llnl

Changes since last release:

• Added a debug patch to help troubleshoot hang in zfetch on Lustre OSTs

zfs-0.7.11-5llnl

No change, just bump version number for new kernel.

Note that the rare incremental send/recv file corruption bug from 0.7.11-4llnl is still present in this release. See the https://github.com/LLNL/zfs/releases/tag/zfs-0.7.11-4llnl release notes for details.

zfs-0.7.11-4llnl

• Revert "Fix problems receiving reallocated dnodes" (d2c8103)
• Revert "Fix object reclaim when using large dnodes" (dc3eea8)
• Revert "Fix 'zfs recv' of non large_dnode send streams" (45f0437)
• Revert "Fix race in dnode_check_slots_free()" (0a8f4e9)

This should fix the "list_del corruption" bug we were seeing (for real this time). However, reverting these patches also exposes a rare bug with incremental send/receives. You may experience file corruption if you do an incremental send/receive of a dataset with varying dnode size. If your dnode size is constant, or you do a full send receive, you will not hit the bug. See openzfs#7617

zfs-0.7.11-3llnl

• Fix arc_release() refcount openzfs#8000 (openzfs@5e8ff25)
• Fix race in dnode_check_slots_free() (openzfs@4764f6f)

This should fix the "list_del corruption" bug we were seeing.

zfs-0.7.11-2llnl

Bump revision for Lustre timestruc_t SPL compatibility fix (LLNL/spl@bf00ca6).

zfs-0.7.11-1llnl

ZFS Version 0.7.11

zfs-0.7.8-4llnl

No changes, just bumped numbering to "4llnl" to make koji happy.

zfs-0.7.8-3llnl

Added updated version of:

• RHEL 7.5 compat: FMODE_KABI_ITERATE openzfs#7460 openzfs#7463