zfs/include/sys
Christian Schwarz 63f8a472af Shared L2ARC - Proof of Concept
(I will give a talk on this PoC at the OpenZFS Developer Summit 2022.)

The ARC dynamically shares DRAM capacity among all currently imported zpools.
However, the L2ARC does not do the same for block capacity: the L2ARC vdevs of
one zpool only cache buffers of that zpool. This can be undesirable on systems
that host multiple zpools because it inhibits dynamic sharing of the cache
device capacity which is desirable if the need for L2ARC varies among zpools
over time, or if the set of zpools that are imported in the system varies over
time.

Shared L2ARC addresses this need by decoupling the L2ARC vdevs from the
zpools that store actual data. The mechanism that we use is to place the L2ARC
vdevs into a special zpool, and to adjust the L2ARC feed thread logic to use
that special zpool's L2ARC vdevs for all zpools' buffers.

High-level changes:

* Reserve "NTNX-fsvm-local-l2arc" as a magic zpool name.
  We call this "the l2arc pool".
  All other pools are called "primary pools".
* Make l2arc feed thread feed ARC buffers from any zpool to the l2arc zpool.
  (Before this patch, the l2arc feed thread would only feed ARC buffers to
  l2arc devices if they are for the same spa_t).
* Change the locking to ensure that the l2arc zpool cannot be removed while
  there are ongoing reads initiated by arc_read on one of the primary pools.

This is sufficient and retains correctness of the ARC because nothing
about the fundamental operation of L2ARC changes. The only thing that changes
is that the L2ARC data is stored on vdevs outside the primary pool.

Proof Of Concept => Production
==============================

This commit is a proof-of-concept.
It works, it results in the desired performance improvement, and it's stable.
But to make it production ready, more work needs to be done.

(1) The design is based on a version of ZFS that does not support
encryption nor Persisent L2ARC. I'm no expert in either of these features.
Encryption might work just fine as long as the l2arc feed thread can access
the encryption keys for l2arc_apply_transforms.
But Persistent L2ARC definitely needs more design work
(multiple L2ARC headers?).

(2) Remove hard-coded magic name; use a property instead.
Make it opt-in so that existing setups are not affected.
Example:
  zpool create -o share_l2arc_vdevs=on my-l2arc-pool

(3) Coexistence with non-shared L2ARC; also via property.
Make it opt-in so that existing setups are not affected.
Example:
  zpool set use_shared_l2arc=on my-data-pool

Signed-off-by: Christian Schwarz <christian.schwarz@nutanix.com>
2022-10-20 11:31:59 +00:00
..
crypto Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
fm Cleanup: Specify unsignedness on things that should not be signed 2022-09-27 16:42:41 -07:00
fs Shared L2ARC - Proof of Concept 2022-10-20 11:31:59 +00:00
lua autoconf: single-step includes 2022-05-10 10:18:51 -07:00
sysevent zed: Fix config_sync autoexpand flood 2022-09-08 10:32:30 -07:00
zstd Unbreak zstd build on sparc64 2022-05-25 09:18:49 -07:00
abd.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
abd_impl.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
aggsum.h More aggsum optimizations 2021-06-07 09:02:47 -07:00
arc.h Shared L2ARC - Proof of Concept 2022-10-20 11:31:59 +00:00
arc_impl.h Cleanup: Specify unsignedness on things that should not be signed 2022-09-27 16:42:41 -07:00
avl.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
avl_impl.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
bitops.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
blake3.h Fix BLAKE3 tuneable and module loading on Linux and FreeBSD 2022-09-16 14:25:53 -07:00
blkptr.h OpenZFS 8067 - zdb should be able to dump literal embedded block pointer 2017-07-07 11:28:01 -07:00
bplist.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
bpobj.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
bptree.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
bqueue.h zfs recv hangs if max recordsize is less than received recordsize 2022-09-16 13:52:25 -07:00
btree.h Several B-tree optimizations 2022-06-24 13:55:58 -07:00
dataset_kstats.h Add support for per dataset zil stats and use wmsum counters 2022-07-20 17:14:06 -07:00
dbuf.h Dynamically size dbuf hash mutex array 2022-09-22 12:59:56 -07:00
ddt.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
dmu.h Fix double const qualifier declarations 2022-09-30 15:34:39 -07:00
dmu_impl.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
dmu_objset.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
dmu_recv.h Implement a new type of zfs receive: corrective receive (-c) 2022-07-28 15:52:46 -07:00
dmu_redact.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
dmu_send.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
dmu_traverse.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
dmu_tx.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
dmu_zfetch.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
dnode.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
dsl_bookmark.h Remaining {=> const} char|void *tag 2022-06-29 14:08:59 -07:00
dsl_crypt.h Fix zpool status in case of unloaded keys 2022-08-22 17:42:01 -07:00
dsl_dataset.h Add zfs.sync.snapshot_rename 2022-09-02 13:31:19 -07:00
dsl_deadlist.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
dsl_deleg.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
dsl_destroy.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
dsl_dir.h Add DD_FIELD string for snapshots_changed property 2022-09-02 13:33:50 -07:00
dsl_pool.h Cleanup: Specify unsignedness on things that should not be signed 2022-09-27 16:42:41 -07:00
dsl_prop.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
dsl_scan.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
dsl_synctask.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
dsl_userhold.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
edonr.h OpenZFS 4185 - add new cryptographic checksums to ZFS: SHA-512, Skein, Edon-R 2016-10-03 14:51:15 -07:00
efi_partition.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
frame.h Linux 5.10 compat: frame.h renamed objtool.h 2020-11-02 22:01:10 +00:00
hkdf.h Encryption patch follow-up 2017-10-11 16:54:48 -04:00
metaslab.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
metaslab_impl.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
mmp.h Add zfs_multihost_interval tunable handler for FreeBSD 2020-06-23 13:32:42 -07:00
mntent.h Expose ZFS dataset case sensitivity setting via sb_opts 2022-07-14 10:38:16 -07:00
mod.h linux: module: weld all but spl.ko into zfs.ko 2022-04-20 13:28:24 -07:00
multilist.h Re-embed multilist_t storage 2021-06-10 10:42:31 -06:00
nvpair.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
nvpair_impl.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
objlist.h Implement Redacted Send/Receive 2019-06-19 09:48:12 -07:00
pathname.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
qat.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
range_tree.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
rrwlock.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
sa.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
sa_impl.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
skein.h OpenZFS 4185 - add new cryptographic checksums to ZFS: SHA-512, Skein, Edon-R 2016-10-03 14:51:15 -07:00
spa.h zed: mark disks as REMOVED when they are removed 2022-09-28 09:48:46 -07:00
spa_checkpoint.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
spa_checksum.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
spa_impl.h Cleanup: Specify unsignedness on things that should not be signed 2022-09-27 16:42:41 -07:00
spa_log_spacemap.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
space_map.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
space_reftree.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
sysevent.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
txg.h Cleanup: Specify unsignedness on things that should not be signed 2022-09-27 16:42:41 -07:00
txg_impl.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
u8_textprep.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
u8_textprep_data.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
uberblock.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
uberblock_impl.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
uio_impl.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
unique.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
uuid.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
vdev.h zed: mark disks as REMOVED when they are removed 2022-09-28 09:48:46 -07:00
vdev_disk.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
vdev_draid.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
vdev_file.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
vdev_impl.h zed: mark disks as REMOVED when they are removed 2022-09-28 09:48:46 -07:00
vdev_indirect_births.h OpenZFS 7614, 9064 - zfs device evacuation/removal 2018-04-14 12:16:17 -07:00
vdev_indirect_mapping.h OpenZFS 7614, 9064 - zfs device evacuation/removal 2018-04-14 12:16:17 -07:00
vdev_initialize.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
vdev_raidz.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
vdev_raidz_impl.h Cleanup Raid-Z Typo fixes 2022-09-06 09:43:21 -07:00
vdev_rebuild.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
vdev_removal.h Cleanup: Specify unsignedness on things that should not be signed 2022-09-27 16:42:41 -07:00
vdev_trim.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
xvattr.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
zap.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
zap_impl.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
zap_leaf.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
zcp.h Remaining {=> const} char|void *tag 2022-06-29 14:08:59 -07:00
zcp_global.h OpenZFS 7431 - ZFS Channel Programs 2018-02-08 15:28:18 -08:00
zcp_iter.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
zcp_prop.h OpenZFS 7431 - ZFS Channel Programs 2018-02-08 15:28:18 -08:00
zcp_set.h Support setting user properties in a channel program 2020-02-14 13:41:42 -08:00
zfeature.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
zfs_acl.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
zfs_bootenv.h zfs label bootenv should store data as nvlist 2020-09-15 15:42:27 -07:00
zfs_chksum.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
zfs_context.h Reduce false positives from Static Analyzers 2022-09-30 15:30:12 -07:00
zfs_debug.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
zfs_delay.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
zfs_file.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
zfs_fuid.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
zfs_ioctl.h DMU_BACKUP_FEATURE: indicate that bit 28 and 29 are reserved 2022-09-27 16:55:32 -07:00
zfs_ioctl_impl.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
zfs_onexit.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
zfs_project.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
zfs_quota.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
zfs_racct.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
zfs_ratelimit.h Change checksum & IO delay ratelimit values 2018-03-04 17:34:51 -08:00
zfs_refcount.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
zfs_rlock.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
zfs_sa.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
zfs_stat.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
zfs_sysfs.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
zfs_vfsops.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
zfs_vnops.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
zfs_znode.h zfs_enter rework 2022-09-16 13:36:47 -07:00
zil.h Add support for per dataset zil stats and use wmsum counters 2022-07-20 17:14:06 -07:00
zil_impl.h Add support for per dataset zil stats and use wmsum counters 2022-07-20 17:14:06 -07:00
zio.h Implement a new type of zfs receive: corrective receive (-c) 2022-07-28 15:52:46 -07:00
zio_checksum.h Fix double const qualifier declarations 2022-09-30 15:34:39 -07:00
zio_compress.h Fix double const qualifier declarations 2022-09-30 15:34:39 -07:00
zio_crypt.h Enable -Wwrite-strings 2022-06-29 14:08:54 -07:00
zio_impl.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
zio_priority.h Add device rebuild feature 2020-07-03 11:05:50 -07:00
zrlock.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
zthr.h Avoid memory allocations in the ARC eviction thread 2022-01-21 10:28:13 -08:00
zvol.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00
zvol_impl.h Replace dead opensolaris.org license link 2022-07-11 14:16:13 -07:00