> Any chance this machine is running many instances of Oracle? I've seen > very high kstat usage in that scenario. Yes: ~75 instances in 6 zones).
> Next step would be to add a '-s 20' to the lockstat to get a kernel > stack for this lock. The output: Adaptive mutex spin: 113576 events in 11.110 seconds (10223 events/sec) ------------------------------------------------------------------------------- Count indv cuml rcnt nsec Lock Caller 1313 1% 1% 0.00 42142202 kstat_chain_lock kstat_hold+0x10 nsec ------ Time Distribution ------ count Stack 1024 | 1 kstat_hold_bykid+0x18 2048 | 13 read_kstat_data+0xa4 4096 |@ 57 fop_ioctl+0x20 8192 |@@ 124 ioctl+0x184 16384 |@@@ 149 syscall_trap+0xac 32768 |@@ 119 65536 |@@ 117 131072 |@@ 123 262144 | 40 524288 | 34 1048576 | 8 2097152 | 13 4194304 | 16 8388608 |@ 49 16777216 |@ 59 33554432 |@ 65 67108864 |@@ 120 134217728 |@ 64 268435456 |@ 48 536870912 |@ 55 1073741824 | 13 ------------------------------------------------------------------------------- Count indv cuml rcnt nsec Lock Caller 1094 1% 2% 0.00 14656339 kstat_chain_lock kstat_rele+0xc nsec ------ Time Distribution ------ count Stack 2048 | 18 read_kstat_data+0x244 4096 |@@ 76 fop_ioctl+0x20 8192 |@@@@ 151 ioctl+0x184 16384 |@@@@@ 208 syscall_trap+0xac 32768 |@@@@ 150 65536 |@@@@ 149 131072 |@@@ 114 262144 |@ 55 524288 | 29 1048576 | 14 2097152 | 13 4194304 | 7 8388608 | 3 16777216 | 22 33554432 | 7 67108864 | 19 134217728 | 14 268435456 | 23 536870912 | 16 1073741824 | 5 ------------------------------------------------------------------------------- Count indv cuml rcnt nsec Lock Caller 735 1% 3% 0.00 10435 0x1992470 page_get_mnode_freelist+0x150 nsec ------ Time Distribution ------ count Stack 2048 | 12 page_get_replacement_page+0x310 4096 |@@@@@@@@@@ 249 page_claim_contig_pages+0x178 8192 |@@@@@@@ 182 page_geti_contig_pages+0x614 16384 |@@@@@@ 159 page_get_contig_pages+0x160 32768 |@@@ 77 page_get_freelist+0x430 65536 |@ 25 page_alloc_pages+0x110 131072 | 4 anon_map_privatepages+0xa4 262144 | 1 anon_map_getpages+0xaa4 524288 | 1 segvn_fault_anonpages+0x384 segvn_fault+0x530 as_fault+0x4c8 pagefault+0xac trap+0xd50 utl0+0x4c ------------------------------------------------------------------------------- Count indv cuml rcnt nsec Lock Caller 727 1% 3% 0.00 9716 0x1992a10 page_get_mnode_freelist+0x150 nsec ------ Time Distribution ------ count Stack 2048 | 3 page_get_replacement_page+0x310 4096 |@@@@@@@@@ 234 page_claim_contig_pages+0x178 8192 |@@@@@@@@ 211 page_geti_contig_pages+0x614 16384 |@@@@@@ 164 page_get_contig_pages+0x160 32768 |@@@ 75 page_get_freelist+0x430 65536 | 21 page_alloc_pages+0x110 131072 | 5 anon_map_privatepages+0xa4 anon_map_getpages+0xaa4 segvn_fault_anonpages+0x384 segvn_fault+0x530 as_fault+0x4c8 pagefault+0xac trap+0xd50 utl0+0x4c ------------------------------------------------------------------------------- Count indv cuml rcnt nsec Lock Caller 717 1% 4% 0.00 239653 vx_worklist_lk vx_worklist_process+0x98 nsec ------ Time Distribution ------ count Stack 2048 | 7 vx_worklist_thread+0x94 4096 |@@@ 81 thread_start+0x4 8192 |@@@@@@ 155 16384 |@@@@@ 120 32768 |@@@@ 107 65536 |@@@ 83 131072 |@ 41 262144 |@ 30 524288 | 12 1048576 | 22 2097152 | 21 4194304 | 15 8388608 | 8 16777216 | 1 ------------------------------------------------------------------------------- [...] Thanks a lot for your help. Matt -- This message posted from opensolaris.org _______________________________________________ perf-discuss mailing list perf-discuss@opensolaris.org