Add L2ARC arcstats for MFU/MRU buffers and buffer content type

Currently the ARC state (MFU/MRU) of cached L2ARC buffer and their
content type is unknown. Knowing this information  may prove beneficial
in adjusting the L2ARC caching policy.

This commit adds L2ARC arcstats that display the aligned size
(in bytes) of L2ARC buffers according to their content type
(data/metadata) and according to their ARC state (MRU/MFU or
prefetch). It also expands the existing evict_l2_eligible arcstat to
differentiate between MFU and MRU buffers.

L2ARC caches buffers from the MRU and MFU lists of ARC. Upon caching a
buffer, its ARC state (MRU/MFU) is stored in the L2 header
(b_arcs_state). The l2_m{f,r}u_asize arcstats reflect the aligned size
(in bytes) of L2ARC buffers according to their ARC state (based on
b_arcs_state). We also account for the case where an L2ARC and ARC
cached MRU or MRU_ghost buffer transitions to MFU. The l2_prefetch_asize
reflects the alinged size (in bytes) of L2ARC buffers that were cached
while they had the prefetch flag set in ARC. This is dynamically updated
as the prefetch flag of L2ARC buffers changes.

When buffers are evicted from ARC, if they are determined to be L2ARC
eligible then their logical size is recorded in
evict_l2_eligible_m{r,f}u arcstats according to their ARC state upon
eviction.

Persistent L2ARC:
When commiting an L2ARC buffer to a log block (L2ARC metadata) its
b_arcs_state and prefetch flag is also stored. If the buffer changes
its arcstate or prefetch flag this is reflected in the above arcstats.
However, the L2ARC metadata cannot currently be updated to reflect this
change.
Example: L2ARC caches an MRU buffer. L2ARC metadata and arcstats count
this as an MRU buffer. The buffer transitions to MFU. The arcstats are
updated to reflect this. Upon pool re-import or on/offlining the L2ARC
device the arcstats are cleared and the buffer will now be counted as an
MRU buffer, as the L2ARC metadata were not updated.

Bug fix:
- If l2arc_noprefetch is set, arc_read_done clears the L2CACHE flag of
  an ARC buffer. However, prefetches may be issued in a way that
  arc_read_done() is bypassed. Instead, move the related code in
  l2arc_write_eligible() to account for those cases too.

Also add a test and update manpages for l2arc_mfuonly module parameter,
and update the manpages and code comments for l2arc_noprefetch.
Move persist_l2arc tests to l2arc.

Signed-off-by: George Amanakis <[email protected]>

Author: gamanakis

cmd/arc_summary/arc_summary2

@@ -219,12 +219,30 @@ def get_arc_summary(Kstat):
     deleted = Kstat["kstat.zfs.misc.arcstats.deleted"]
     mutex_miss = Kstat["kstat.zfs.misc.arcstats.mutex_miss"]
     evict_skip = Kstat["kstat.zfs.misc.arcstats.evict_skip"]
+    evict_l2_cached = Kstat["kstat.zfs.misc.arcstats.evict_l2_cached"]
+    evict_l2_eligible = Kstat["kstat.zfs.misc.arcstats.evict_l2_eligible"]
+    evict_l2_eligible_mfu = Kstat["kstat.zfs.misc.arcstats.evict_l2_eligible_mfu"]
+    evict_l2_eligible_mru = Kstat["kstat.zfs.misc.arcstats.evict_l2_eligible_mru"]
+    evict_l2_ineligible = Kstat["kstat.zfs.misc.arcstats.evict_l2_ineligible"]
+    evict_l2_skip = Kstat["kstat.zfs.misc.arcstats.evict_l2_skip"]
 
     # ARC Misc.
     output["arc_misc"] = {}
     output["arc_misc"]["deleted"] = fHits(deleted)
-    output["arc_misc"]['mutex_miss'] = fHits(mutex_miss)
-    output["arc_misc"]['evict_skips'] = fHits(evict_skip)
+    output["arc_misc"]["mutex_miss"] = fHits(mutex_miss)
+    output["arc_misc"]["evict_skips"] = fHits(evict_skip)
+    output["arc_misc"]["evict_l2_skip"] = fHits(evict_l2_skip)
+    output["arc_misc"]["evict_l2_cached"] = fBytes(evict_l2_cached)
+    output["arc_misc"]["evict_l2_eligible"] = fBytes(evict_l2_eligible)
+    output["arc_misc"]["evict_l2_eligible_mfu"] = {
+            'per': fPerc(evict_l2_eligible_mfu, evict_l2_eligible),
+            'num': fBytes(evict_l2_eligible_mfu),
+    }
+    output["arc_misc"]["evict_l2_eligible_mru"] = {
+            'per': fPerc(evict_l2_eligible_mru, evict_l2_eligible),
+            'num': fBytes(evict_l2_eligible_mru),
+    }
+    output["arc_misc"]["evict_l2_ineligible"] = fBytes(evict_l2_ineligible)
 
     # ARC Sizing
     arc_size = Kstat["kstat.zfs.misc.arcstats.size"]
@@ -340,8 +358,26 @@ def _arc_summary(Kstat):
     sys.stdout.write("\tDeleted:\t\t\t\t%s\n" % arc['arc_misc']['deleted'])
     sys.stdout.write("\tMutex Misses:\t\t\t\t%s\n" %
                      arc['arc_misc']['mutex_miss'])
-    sys.stdout.write("\tEvict Skips:\t\t\t\t%s\n" %
+    sys.stdout.write("\tEviction Skips:\t\t\t\t%s\n" %
                      arc['arc_misc']['evict_skips'])
+    sys.stdout.write("\tEviction Skips Due to L2 Writes:\t%s\n" %
+                     arc['arc_misc']['evict_l2_skip'])
+    sys.stdout.write("\tL2 Cached Evictions:\t\t\t%s\n" %
+                     arc['arc_misc']['evict_l2_cached'])
+    sys.stdout.write("\tL2 Eligible Evictions:\t\t\t%s\n" %
+                     arc['arc_misc']['evict_l2_eligible'])
+    sys.stdout.write("\tL2 Eligible MFU Evictions:\t%s\t%s\n" % (
+                     arc['arc_misc']['evict_l2_eligible_mfu']['per'],
+                     arc['arc_misc']['evict_l2_eligible_mfu']['num'],
+                     )
+    )
+    sys.stdout.write("\tL2 Eligible MRU Evictions:\t%s\t%s\n" % (
+                     arc['arc_misc']['evict_l2_eligible_mru']['per'],
+                     arc['arc_misc']['evict_l2_eligible_mru']['num'],
+                     )
+    )
+    sys.stdout.write("\tL2 Ineligible Evictions:\t\t%s\n" %
+                     arc['arc_misc']['evict_l2_ineligible'])
     sys.stdout.write("\n")
 
     # ARC Sizing
@@ -677,6 +713,11 @@ def get_l2arc_summary(Kstat):
     l2_writes_done = Kstat["kstat.zfs.misc.arcstats.l2_writes_done"]
     l2_writes_error = Kstat["kstat.zfs.misc.arcstats.l2_writes_error"]
     l2_writes_sent = Kstat["kstat.zfs.misc.arcstats.l2_writes_sent"]
+    l2_mfu_asize = Kstat["kstat.zfs.misc.arcstats.l2_mfu_asize"]
+    l2_mru_asize = Kstat["kstat.zfs.misc.arcstats.l2_mru_asize"]
+    l2_prefetch_asize = Kstat["kstat.zfs.misc.arcstats.l2_prefetch_asize"]
+    l2_bufc_data_asize = Kstat["kstat.zfs.misc.arcstats.l2_bufc_data_asize"]
+    l2_bufc_metadata_asize = Kstat["kstat.zfs.misc.arcstats.l2_bufc_metadata_asize"]
 
     l2_access_total = (l2_hits + l2_misses)
     output['l2_health_count'] = (l2_writes_error + l2_cksum_bad + l2_io_error)
@@ -699,7 +740,7 @@ def get_l2arc_summary(Kstat):
         output["io_errors"] = fHits(l2_io_error)
 
         output["l2_arc_size"] = {}
-        output["l2_arc_size"]["adative"] = fBytes(l2_size)
+        output["l2_arc_size"]["adaptive"] = fBytes(l2_size)
         output["l2_arc_size"]["actual"] = {
             'per': fPerc(l2_asize, l2_size),
             'num': fBytes(l2_asize)
@@ -708,6 +749,26 @@ def get_l2arc_summary(Kstat):
             'per': fPerc(l2_hdr_size, l2_size),
             'num': fBytes(l2_hdr_size),
         }
+        output["l2_arc_size"]["mfu_asize"] = {
+            'per': fPerc(l2_mfu_asize, l2_asize),
+            'num': fBytes(l2_mfu_asize),
+        }
+        output["l2_arc_size"]["mru_asize"] = {
+            'per': fPerc(l2_mru_asize, l2_asize),
+            'num': fBytes(l2_mru_asize),
+        }
+        output["l2_arc_size"]["prefetch_asize"] = {
+            'per': fPerc(l2_prefetch_asize, l2_asize),
+            'num': fBytes(l2_prefetch_asize),
+        }
+        output["l2_arc_size"]["bufc_data_asize"] = {
+            'per': fPerc(l2_bufc_data_asize, l2_asize),
+            'num': fBytes(l2_bufc_data_asize),
+        }
+        output["l2_arc_size"]["bufc_metadata_asize"] = {
+            'per': fPerc(l2_bufc_metadata_asize, l2_asize),
+            'num': fBytes(l2_bufc_metadata_asize),
+        }
 
         output["l2_arc_evicts"] = {}
         output["l2_arc_evicts"]['lock_retries'] = fHits(l2_evict_lock_retry)
@@ -772,7 +833,7 @@ def _l2arc_summary(Kstat):
         sys.stdout.write("\n")
 
         sys.stdout.write("L2 ARC Size: (Adaptive)\t\t\t\t%s\n" %
-                         arc["l2_arc_size"]["adative"])
+                         arc["l2_arc_size"]["adaptive"])
         sys.stdout.write("\tCompressed:\t\t\t%s\t%s\n" % (
             arc["l2_arc_size"]["actual"]["per"],
             arc["l2_arc_size"]["actual"]["num"],
@@ -783,11 +844,36 @@ def _l2arc_summary(Kstat):
             arc["l2_arc_size"]["head_size"]["num"],
             )
         )
+        sys.stdout.write("\tMFU Alloc. Size:\t\t%s\t%s\n" % (
+            arc["l2_arc_size"]["mfu_asize"]["per"],
+            arc["l2_arc_size"]["mfu_asize"]["num"],
+            )
+        )
+        sys.stdout.write("\tMRU Alloc. Size:\t\t%s\t%s\n" % (
+            arc["l2_arc_size"]["mru_asize"]["per"],
+            arc["l2_arc_size"]["mru_asize"]["num"],
+            )
+        )
+        sys.stdout.write("\tPrefetch Alloc. Size:\t\t%s\t%s\n" % (
+            arc["l2_arc_size"]["prefetch_asize"]["per"],
+            arc["l2_arc_size"]["prefetch_asize"]["num"],
+            )
+        )
+        sys.stdout.write("\tData (buf content) Alloc. Size:\t%s\t%s\n" % (
+            arc["l2_arc_size"]["bufc_data_asize"]["per"],
+            arc["l2_arc_size"]["bufc_data_asize"]["num"],
+            )
+        )
+        sys.stdout.write("\tMetadata (buf content) Size:\t%s\t%s\n" % (
+            arc["l2_arc_size"]["bufc_metadata_asize"]["per"],
+            arc["l2_arc_size"]["bufc_metadata_asize"]["num"],
+            )
+        )
         sys.stdout.write("\n")
 
         if arc["l2_arc_evicts"]['lock_retries'] != '0' or \
            arc["l2_arc_evicts"]["reading"] != '0':
-            sys.stdout.write("L2 ARC Evicts:\n")
+            sys.stdout.write("L2 ARC Evictions:\n")
             sys.stdout.write("\tLock Retries:\t\t\t\t%s\n" %
                              arc["l2_arc_evicts"]['lock_retries'])
             sys.stdout.write("\tUpon Reading:\t\t\t\t%s\n" %

cmd/arc_summary/arc_summary3

@@ -598,6 +598,20 @@ def section_arc(kstats_dict):
     prt_i1('Deleted:', f_hits(arc_stats['deleted']))
     prt_i1('Mutex misses:', f_hits(arc_stats['mutex_miss']))
     prt_i1('Eviction skips:', f_hits(arc_stats['evict_skip']))
+    prt_i1('Eviction skips due to L2 writes:',
+           f_hits(arc_stats['evict_l2_skip']))
+    prt_i1('L2 cached evictions:', f_bytes(arc_stats['evict_l2_cached']))
+    prt_i1('L2 eligible evictions:', f_bytes(arc_stats['evict_l2_eligible']))
+    prt_i2('L2 eligible MFU evictions:',
+           f_perc(arc_stats['evict_l2_eligible_mfu'],
+           arc_stats['evict_l2_eligible']),
+           f_bytes(arc_stats['evict_l2_eligible_mfu']))
+    prt_i2('L2 eligible MRU evictions:',
+           f_perc(arc_stats['evict_l2_eligible_mru'],
+           arc_stats['evict_l2_eligible']),
+           f_bytes(arc_stats['evict_l2_eligible_mru']))
+    prt_i1('L2 ineligible evictions:',
+           f_bytes(arc_stats['evict_l2_ineligible']))
     print()
 
 
@@ -736,6 +750,21 @@ def section_l2arc(kstats_dict):
     prt_i2('Header size:',
            f_perc(arc_stats['l2_hdr_size'], arc_stats['l2_size']),
            f_bytes(arc_stats['l2_hdr_size']))
+    prt_i2('MFU allocated size:',
+           f_perc(arc_stats['l2_mfu_asize'], arc_stats['l2_asize']),
+           f_bytes(arc_stats['l2_mfu_asize']))
+    prt_i2('MRU allocated size:',
+           f_perc(arc_stats['l2_mru_asize'], arc_stats['l2_asize']),
+           f_bytes(arc_stats['l2_mru_asize']))
+    prt_i2('Prefetch allocated size:',
+           f_perc(arc_stats['l2_prefetch_asize'], arc_stats['l2_asize']),
+           f_bytes(arc_stats['l2_prefetch_asize']))
+    prt_i2('Data (buffer content) allocated size:',
+           f_perc(arc_stats['l2_bufc_data_asize'], arc_stats['l2_asize']),
+           f_bytes(arc_stats['l2_bufc_data_asize']))
+    prt_i2('Metadata (buffer content) allocated size:',
+           f_perc(arc_stats['l2_bufc_metadata_asize'], arc_stats['l2_asize']),
+           f_bytes(arc_stats['l2_bufc_metadata_asize']))
 
     print()
     prt_1('L2ARC breakdown:', f_hits(l2_access_total))

cmd/arcstat/arcstat.in

@@ -88,6 +88,12 @@ cols = {
     "mfug":       [4, 1000, "MFU ghost list hits per second"],
     "mrug":       [4, 1000, "MRU ghost list hits per second"],
     "eskip":      [5, 1000, "evict_skip per second"],
+    "el2skip":    [7, 1000, "evict skip, due to l2 writes, per second"],
+    "el2cach":    [7, 1024, "Size of L2 cached evictions per second"],
+    "el2el":      [5, 1024, "Size of L2 eligible evictions per second"],
+    "el2mfu":     [6, 1024, "Size of L2 eligible MFU evictions per second"],
+    "el2mru":     [6, 1024, "Size of L2 eligible MRU evictions per second"],
+    "el2inel":    [7, 1024, "Size of L2 ineligible evictions per second"],
     "mtxmis":     [6, 1000, "mutex_miss per second"],
     "dread":      [5, 1000, "Demand accesses per second"],
     "pread":      [5, 1000, "Prefetch accesses per second"],
@@ -96,6 +102,16 @@ cols = {
     "l2read":     [6, 1000, "Total L2ARC accesses per second"],
     "l2hit%":     [6, 100, "L2ARC access hit percentage"],
     "l2miss%":    [7, 100, "L2ARC access miss percentage"],
+    "l2pref":     [6, 1024, "L2ARC prefetch allocated size"],
+    "l2mfu":      [5, 1024, "L2ARC MFU allocated size"],
+    "l2mru":      [5, 1024, "L2ARC MRU allocated size"],
+    "l2data":     [6, 1024, "L2ARC data allocated size"],
+    "l2meta":     [6, 1024, "L2ARC metadata allocated size"],
+    "l2pref%":    [7, 100, "L2ARC prefetch percentage"],
+    "l2mfu%":     [6, 100, "L2ARC MFU percentage"],
+    "l2mru%":     [6, 100, "L2ARC MRU percentage"],
+    "l2data%":    [7, 100, "L2ARC data percentage"],
+    "l2meta%":    [7, 100, "L2ARC metadata percentage"],
     "l2asize":    [7, 1024, "Actual (compressed) size of the L2ARC"],
     "l2size":     [6, 1024, "Size of the L2ARC"],
     "l2bytes":    [7, 1024, "Bytes read per second from the L2ARC"],
@@ -436,6 +452,12 @@ def calculate():
     v["mrug"] = d["mru_ghost_hits"] / sint
     v["mfug"] = d["mfu_ghost_hits"] / sint
     v["eskip"] = d["evict_skip"] / sint
+    v["el2skip"] = d["evict_l2_skip"] / sint
+    v["el2cach"] = d["evict_l2_cached"] / sint
+    v["el2el"] = d["evict_l2_eligible"] / sint
+    v["el2mfu"] = d["evict_l2_eligible_mfu"] / sint
+    v["el2mru"] = d["evict_l2_eligible_mru"] / sint
+    v["el2inel"] = d["evict_l2_ineligible"] / sint
     v["mtxmis"] = d["mutex_miss"] / sint
 
     if l2exist:
@@ -449,6 +471,17 @@ def calculate():
         v["l2size"] = cur["l2_size"]
         v["l2bytes"] = d["l2_read_bytes"] / sint
 
+        v["l2pref"] = cur["l2_prefetch_asize"]
+        v["l2mfu"] = cur["l2_mfu_asize"]
+        v["l2mru"] = cur["l2_mru_asize"]
+        v["l2data"] = cur["l2_bufc_data_asize"]
+        v["l2meta"] = cur["l2_bufc_metadata_asize"]
+        v["l2pref%"] = 100 * v["l2pref"] / v["l2asize"]
+        v["l2mfu%"] = 100 * v["l2mfu"] / v["l2asize"]
+        v["l2mru%"] = 100 * v["l2mru"] / v["l2asize"]
+        v["l2data%"] = 100 * v["l2data"] / v["l2asize"]
+        v["l2meta%"] = 100 * v["l2meta"] / v["l2asize"]
+
     v["grow"] = 0 if cur["arc_no_grow"] else 1
     v["need"] = cur["arc_need_free"]
     v["free"] = cur["memory_free_bytes"]

cmd/zdb/zdb.c

@@ -4188,6 +4188,8 @@ dump_l2arc_log_entries(uint64_t log_entries,
 		    (u_longlong_t)L2BLK_GET_PREFETCH((&le[j])->le_prop));
 		(void) printf("|\t\t\t\taddress: %llu\n",
 		    (u_longlong_t)le[j].le_daddr);
+		(void) printf("|\t\t\t\tARC state: %llu\n",
+		    (u_longlong_t)L2BLK_GET_STATE((&le[j])->le_prop));
 		(void) printf("|\n");
 	}
 	(void) printf("\n");

configure.ac

@@ -338,6 +338,7 @@ AC_CONFIG_FILES([
 	tests/zfs-tests/tests/functional/inheritance/Makefile
 	tests/zfs-tests/tests/functional/inuse/Makefile
 	tests/zfs-tests/tests/functional/io/Makefile
+	tests/zfs-tests/tests/functional/l2arc/Makefile
 	tests/zfs-tests/tests/functional/large_files/Makefile
 	tests/zfs-tests/tests/functional/largest_pool/Makefile
 	tests/zfs-tests/tests/functional/libzfs/Makefile
@@ -354,7 +355,6 @@ AC_CONFIG_FILES([
 	tests/zfs-tests/tests/functional/nopwrite/Makefile
 	tests/zfs-tests/tests/functional/online_offline/Makefile
 	tests/zfs-tests/tests/functional/pam/Makefile
-	tests/zfs-tests/tests/functional/persist_l2arc/Makefile
 	tests/zfs-tests/tests/functional/pool_checkpoint/Makefile
 	tests/zfs-tests/tests/functional/pool_names/Makefile
 	tests/zfs-tests/tests/functional/poolversion/Makefile

include/sys/arc_impl.h

@@ -350,6 +350,8 @@ typedef struct l2arc_lb_ptr_buf {
 #define	L2BLK_SET_TYPE(field, x)	BF64_SET((field), 48, 8, x)
 #define	L2BLK_GET_PROTECTED(field)	BF64_GET((field), 56, 1)
 #define	L2BLK_SET_PROTECTED(field, x)	BF64_SET((field), 56, 1, x)
+#define	L2BLK_GET_STATE(field)		BF64_GET((field), 57, 4)
+#define	L2BLK_SET_STATE(field, x)	BF64_SET((field), 57, 4, x)
 
 #define	PTR_SWAP(x, y)		\
 	do {			\
@@ -445,6 +447,7 @@ typedef struct l2arc_buf_hdr {
 	l2arc_dev_t		*b_dev;		/* L2ARC device */
 	uint64_t		b_daddr;	/* disk address, offset byte */
 	uint32_t		b_hits;
+	arc_state_type_t	b_arcs_state;
 	list_node_t		b_l2node;
 } l2arc_buf_hdr_t;
 
@@ -546,6 +549,8 @@ typedef struct arc_stats {
 	kstat_named_t arcstat_evict_not_enough;
 	kstat_named_t arcstat_evict_l2_cached;
 	kstat_named_t arcstat_evict_l2_eligible;
+	kstat_named_t arcstat_evict_l2_eligible_mfu;
+	kstat_named_t arcstat_evict_l2_eligible_mru;
 	kstat_named_t arcstat_evict_l2_ineligible;
 	kstat_named_t arcstat_evict_l2_skip;
 	kstat_named_t arcstat_hash_elements;
@@ -744,6 +749,18 @@ typedef struct arc_stats {
 	kstat_named_t arcstat_mfu_ghost_evictable_metadata;
 	kstat_named_t arcstat_l2_hits;
 	kstat_named_t arcstat_l2_misses;
+	/*
+	 * Allocated size (in bytes) of L2ARC cached buffers by ARC state.
+	 */
+	kstat_named_t arcstat_l2_prefetch_asize;
+	kstat_named_t arcstat_l2_mru_asize;
+	kstat_named_t arcstat_l2_mfu_asize;
+	/*
+	 * Allocated size (in bytes) of L2ARC cached buffers by buffer content
+	 * type.
+	 */
+	kstat_named_t arcstat_l2_bufc_data_asize;
+	kstat_named_t arcstat_l2_bufc_metadata_asize;
 	kstat_named_t arcstat_l2_feeds;
 	kstat_named_t arcstat_l2_rw_clash;
 	kstat_named_t arcstat_l2_read_bytes;