== replay-dual test 26: dbench and tar with mds failover ========================================================== 03:25:30 (1785223530) Starting client oleg405-client.virtnet: -o user_xattr,flock 192.168.204.105@tcp:/lustre /mnt/lustre Started clients oleg405-client.virtnet: 192.168.204.105@tcp:/lustre on /mnt/lustre type lustre (rw,flock,checksum,encrypt,lazystatfs,lruresize,nolock,statfs_project,nosync_on_close,nouser_fid2path,user_xattr,verbose) Started tar loop with pid 65595 Started dbench loop with 65596 striped dir -i0 -c2 -H crush2 /mnt/lustre2/d26.replay-dual/run_dbench striped dir -i0 -c2 -H fnv_1a_64 /mnt/lustre/d26.replay-dual/run_tar looking for dbench program /usr/bin/dbench found dbench client file /usr/share/dbench/client.txt '/usr/share/dbench/client.txt' -> 'client.txt' UUID 1K-blocks Used Available Use% Mounted on running 'dbench 1 -D /mnt/lustre2/d26.replay-dual/run_dbench -t 100' on /mnt/lustre at Tue Jul 28 03:25:34 EDT 2026 waiting for dbench pid 65645 lustre-MDT0000_UUID 1414116 1980 1269324 1% /mnt/lustre[MDT:0] lustre-MDT0001_UUID 1414116 1792 1269512 1% /mnt/lustre[MDT:1] lustre-OST0000_UUID 3833116 1564 3605456 1% /mnt/lustre[OST:0] lustre-OST0001_UUID 3833116 1544 3580680 1% /mnt/lustre[OST:1] filesystem_summary: 7666232 3108 7186136 1% /mnt/lustre dbench version 4.00 - Copyright Andrew Tridgell 1999-2004 Running for 100 seconds with load 'client.txt' and minimum warmup 20 secs failed to create barrier semaphore 0 of 1 processes prepared for launch 0 sec 1 of 1 processes prepared for launch 0 sec releasing clients 1 24 0.65 MB/sec warmup 1 sec latency 11.626 ms 1 97 2.19 MB/sec warmup 2 sec latency 90.981 ms 1 139 2.00 MB/sec warmup 3 sec latency 157.215 ms 1 181 1.83 MB/sec warmup 4 sec latency 136.997 ms 1 204 1.59 MB/sec warmup 5 sec latency 384.203 ms 1 240 1.52 MB/sec warmup 6 sec latency 135.863 ms 1 287 1.53 MB/sec warmup 7 sec latency 117.026 ms 1 329 1.49 MB/sec warmup 8 sec latency 132.636 ms test_26 fail mds1 1 times 1 365 1.45 MB/sec warmup 9 sec latency 128.042 ms 1 421 1.48 MB/sec warmup 10 sec latency 113.400 ms Failing mds1 on oleg405-server 1 462 1.47 MB/sec warmup 11 sec latency 112.009 ms Stopping /mnt/lustre-mds1 (opts:) on oleg405-server 1 511 1.47 MB/sec warmup 12 sec latency 136.032 ms 1 568 1.47 MB/sec warmup 13 sec latency 180.790 ms 1 606 1.50 MB/sec warmup 14 sec latency 586.648 ms 03:25:49 (1785223549) shut down facet: mds1 facet_host: oleg405-server facet_failover_host: oleg405-server 1 606 1.40 MB/sec warmup 15 sec latency 1586.856 ms 1 606 1.31 MB/sec warmup 16 sec latency 2587.961 ms 1 606 1.23 MB/sec warmup 17 sec latency 3588.192 ms 1 606 1.16 MB/sec warmup 18 sec latency 4588.431 ms 1 606 1.10 MB/sec warmup 19 sec latency 5588.667 ms 1 606 0.00 MB/sec execute 1 sec latency 7590.049 ms 1 606 0.00 MB/sec execute 2 sec latency 8590.547 ms 1 606 0.00 MB/sec execute 3 sec latency 9591.130 ms 1 606 0.00 MB/sec execute 4 sec latency 10591.350 ms 1 606 0.00 MB/sec execute 5 sec latency 11591.596 ms Failover mds1 to oleg405-server mount facets: mds1 1 606 0.00 MB/sec execute 6 sec latency 12591.797 ms 1 606 0.00 MB/sec execute 7 sec latency 13592.059 ms 1 606 0.00 MB/sec execute 8 sec latency 14592.319 ms 1 606 0.00 MB/sec execute 9 sec latency 15611.428 ms 1 606 0.00 MB/sec execute 10 sec latency 16611.602 ms 1 606 0.00 MB/sec execute 11 sec latency 17613.100 ms 1 606 0.00 MB/sec execute 12 sec latency 18614.798 ms 1 606 0.00 MB/sec execute 13 sec latency 19615.031 ms 1 606 0.00 MB/sec execute 14 sec latency 20615.183 ms Start mds1: mount -t lustre -o localrecov /dev/mapper/mds1_flakey /mnt/lustre-mds1 1 606 0.00 MB/sec execute 15 sec latency 21617.044 ms 1 606 0.00 MB/sec execute 16 sec latency 22617.666 ms 1 606 0.00 MB/sec execute 17 sec latency 23617.860 ms 1 606 0.00 MB/sec execute 18 sec latency 24619.251 ms 1 606 0.00 MB/sec execute 19 sec latency 25620.052 ms 1 606 0.00 MB/sec execute 20 sec latency 26620.345 ms 1 606 0.00 MB/sec execute 21 sec latency 27620.589 ms 1 606 0.00 MB/sec execute 22 sec latency 28620.755 ms 1 606 0.00 MB/sec execute 23 sec latency 29620.934 ms 1 606 0.00 MB/sec execute 24 sec latency 30621.090 ms 1 606 0.00 MB/sec execute 25 sec latency 31621.316 ms oleg405-server: oleg405-server.virtnet: executing set_default_debug -1 all 1 606 0.00 MB/sec execute 26 sec latency 32622.293 ms 1 606 0.00 MB/sec execute 27 sec latency 33623.311 ms pdsh@oleg405-client: oleg405-server: ssh exited with exit code 1 1 606 0.00 MB/sec execute 28 sec latency 34623.509 ms Started lustre-MDT0000 03:26:23 (1785223583) targets are mounted 03:26:23 (1785223583) facet_failover done 1 606 0.00 MB/sec execute 29 sec latency 35624.414 ms 1 639 0.00 MB/sec execute 30 sec latency 36118.544 ms 1 674 0.01 MB/sec execute 31 sec latency 98.064 ms 1 708 0.01 MB/sec execute 32 sec latency 115.504 ms 1 720 0.01 MB/sec execute 33 sec latency 580.617 ms 1 735 0.01 MB/sec execute 34 sec latency 850.819 ms oleg405-client.virtnet: executing wait_import_state_mount (FULL|IDLE) mdc.lustre-MDT0000-mdc-*.mds_server_uuid 1475 0 1 776 0.01 MB/sec execute 35 sec latency 147.460 ms 1 804 0.01 MB/sec execute 36 sec latency 167.752 ms mdc.lustre-MDT0000-mdc-*.mds_server_uuid in FULL state after 0 sec 1 851 0.02 MB/sec execute 37 sec latency 100.203 ms 1 885 0.02 MB/sec execute 38 sec latency 149.738 ms 1 910 0.02 MB/sec execute 39 sec latency 136.839 ms 1 948 0.02 MB/sec execute 40 sec latency 132.530 ms 1 977 0.02 MB/sec execute 41 sec latency 145.273 ms 1 1028 0.05 MB/sec execute 42 sec latency 103.976 ms 1 1074 0.05 MB/sec execute 43 sec latency 75.483 ms UUID 1K-blocks Used Available Use% Mounted on lustre-MDT0000_UUID 1414116 2192 1269112 1% /mnt/lustre[MDT:0] lustre-MDT0001_UUID 1414116 1956 1269348 1% /mnt/lustre[MDT:1] lustre-OST0000_UUID 3833116 22180 3583576 1% /mnt/lustre[OST:0] lustre-OST0001_UUID 3833116 36808 3569272 2% /mnt/lustre[OST:1] filesystem_summary: 7666232 58988 7152848 1% /mnt/lustre 1 1127 0.05 MB/sec execute 44 sec latency 84.307 ms 1 1161 0.05 MB/sec execute 45 sec latency 144.692 ms 1 1202 0.05 MB/sec execute 46 sec latency 77.655 ms 1 1246 0.04 MB/sec execute 47 sec latency 91.264 ms 1 1319 0.07 MB/sec execute 48 sec latency 232.570 ms 1 1388 0.10 MB/sec execute 49 sec latency 780.905 ms 1 1425 0.10 MB/sec execute 50 sec latency 91.585 ms 1 1474 0.10 MB/sec execute 51 sec latency 106.117 ms 1 1516 0.10 MB/sec execute 52 sec latency 102.711 ms test_26 fail mds2 2 times 1 1572 0.10 MB/sec execute 53 sec latency 76.819 ms 1 1613 0.10 MB/sec execute 54 sec latency 100.478 ms Failing mds2 on oleg405-server 1 1653 0.10 MB/sec execute 55 sec latency 116.523 ms Stopping /mnt/lustre-mds2 (opts:) on oleg405-server 1 1693 0.09 MB/sec execute 56 sec latency 131.718 ms 1 1769 0.09 MB/sec execute 57 sec latency 78.065 ms 1 1818 0.09 MB/sec execute 58 sec latency 119.891 ms 03:26:54 (1785223614) shut down facet: mds2 facet_host: oleg405-server facet_failover_host: oleg405-server 1 1859 0.09 MB/sec execute 59 sec latency 158.076 ms 1 1912 0.09 MB/sec execute 60 sec latency 111.566 ms 1 1989 0.09 MB/sec execute 61 sec latency 75.372 ms 1 2069 0.09 MB/sec execute 62 sec latency 59.830 ms 1 2154 0.09 MB/sec execute 63 sec latency 76.740 ms 1 2244 0.09 MB/sec execute 64 sec latency 59.952 ms 1 2259 0.09 MB/sec execute 65 sec latency 775.756 ms 1 2259 0.09 MB/sec execute 66 sec latency 1777.544 ms 1 2259 0.09 MB/sec execute 67 sec latency 2777.903 ms 1 2259 0.09 MB/sec execute 68 sec latency 3778.160 ms 1 2259 0.08 MB/sec execute 69 sec latency 4778.824 ms Failover mds2 to oleg405-server mount facets: mds2 1 2259 0.08 MB/sec execute 70 sec latency 5779.063 ms 1 2259 0.08 MB/sec execute 71 sec latency 6781.854 ms 1 2259 0.08 MB/sec execute 72 sec latency 7783.175 ms 1 2259 0.08 MB/sec execute 73 sec latency 8783.416 ms 1 2259 0.08 MB/sec execute 74 sec latency 9785.745 ms 1 2259 0.08 MB/sec execute 75 sec latency 10785.949 ms 1 2259 0.08 MB/sec execute 76 sec latency 11786.164 ms 1 2259 0.08 MB/sec execute 77 sec latency 12786.465 ms Start mds2: mount -t lustre -o localrecov /dev/mapper/mds2_flakey /mnt/lustre-mds2 1 2259 0.07 MB/sec execute 78 sec latency 13790.643 ms 1 2259 0.07 MB/sec execute 79 sec latency 14790.801 ms 1 2259 0.07 MB/sec execute 80 sec latency 15791.007 ms 1 2259 0.07 MB/sec execute 81 sec latency 16791.231 ms 1 2259 0.07 MB/sec execute 82 sec latency 17791.388 ms oleg405-server: oleg405-server.virtnet: executing set_default_debug -1 all pdsh@oleg405-client: oleg405-server: ssh exited with exit code 1 1 2259 0.07 MB/sec execute 83 sec latency 18791.766 ms Started lustre-MDT0001 03:27:18 (1785223638) targets are mounted 03:27:18 (1785223638) facet_failover done 1 2259 0.07 MB/sec execute 84 sec latency 19793.046 ms 1 2259 0.07 MB/sec execute 85 sec latency 20795.388 ms 1 2326 0.08 MB/sec execute 86 sec latency 21260.595 ms 1 2369 0.08 MB/sec execute 87 sec latency 68.079 ms 1 2412 0.08 MB/sec execute 88 sec latency 110.406 ms 1 2463 0.08 MB/sec execute 89 sec latency 78.095 ms oleg405-client.virtnet: executing wait_import_state_mount (FULL|IDLE) mdc.lustre-MDT0001-mdc-*.mds_server_uuid 1475 0 1 2554 0.09 MB/sec execute 90 sec latency 108.944 ms mdc.lustre-MDT0001-mdc-*.mds_server_uuid in FULL state after 0 sec 1 2602 0.09 MB/sec execute 91 sec latency 88.898 ms 1 2770 0.12 MB/sec execute 92 sec latency 87.405 ms 1 2915 0.13 MB/sec execute 93 sec latency 57.490 ms 1 2990 0.13 MB/sec execute 94 sec latency 69.421 ms 1 3189 0.14 MB/sec execute 95 sec latency 101.279 ms 1 3291 0.15 MB/sec execute 96 sec latency 99.878 ms UUID 1K-blocks Used Available Use% Mounted on lustre-MDT0000_UUID 1414116 2500 1268804 1% /mnt/lustre[MDT:0] lustre-MDT0001_UUID 1414116 2280 1269024 1% /mnt/lustre[MDT:1] lustre-OST0000_UUID 3833116 31952 3572808 1% /mnt/lustre[OST:0] lustre-OST0001_UUID 3833116 37652 3566268 2% /mnt/lustre[OST:1] filesystem_summary: 7666232 69604 7139076 1% /mnt/lustre 1 3375 0.15 MB/sec execute 97 sec latency 60.783 ms 1 3541 0.18 MB/sec execute 98 sec latency 68.430 ms 1 3627 0.18 MB/sec execute 99 sec latency 69.450 ms 1 cleanup 100 sec 1 cleanup 101 sec 1 cleanup 102 sec 1 cleanup 103 sec 0 cleanup 104 sec Operation Count AvgLat MaxLat ---------------------------------------- NTCreateX 468 90.807 21260.580 Close 391 10.362 42.480 Rename 25 83.817 167.725 Unlink 59 22.939 49.567 Qpathinfo 465 92.538 36118.523 Qfileinfo 95 2.069 24.166 Qfsinfo 86 2.045 21.554 Sfileinfo 50 58.369 126.402 Find 191 9.289 27.753 WriteX 275 11.743 80.766 ReadX 938 1.262 96.653 LockX 2 5.191 5.347 UnlockX 2 5.280 5.302 Flush 43 57.550 192.185 Throughput 0.177842 MB/sec 1 clients 1 procs max_latency=36118.544 ms stopping dbench on /mnt/lustre at Tue Jul 28 03:27:38 EDT 2026 with return code 0 clean dbench files on /mnt/lustre /mnt/lustre /mnt/lustre test_26 fail mds1 3 times removed 'client.txt' /mnt/lustre dbench successfully finished Failing mds1 on oleg405-server striped dir -i0 -c2 -H crush /mnt/lustre2/d26.replay-dual/run_dbench Stopping /mnt/lustre-mds1 (opts:) on oleg405-server 03:27:44 (1785223664) shut down facet: mds1 facet_host: oleg405-server facet_failover_host: oleg405-server Failover mds1 to oleg405-server mount facets: mds1 Start mds1: mount -t lustre -o localrecov /dev/mapper/mds1_flakey /mnt/lustre-mds1 oleg405-server: oleg405-server.virtnet: executing set_default_debug -1 all pdsh@oleg405-client: oleg405-server: ssh exited with exit code 1 Started lustre-MDT0000 03:28:19 (1785223699) targets are mounted 03:28:19 (1785223699) facet_failover done looking for dbench program /usr/bin/dbench found dbench client file /usr/share/dbench/client.txt '/usr/share/dbench/client.txt' -> 'client.txt' running 'dbench 1 -D /mnt/lustre2/d26.replay-dual/run_dbench -t 100' on /mnt/lustre at Tue Jul 28 03:28:24 EDT 2026 waiting for dbench pid 68432 dbench version 4.00 - Copyright Andrew Tridgell 1999-2004 Running for 100 seconds with load 'client.txt' and minimum warmup 20 secs 0 of 1 processes prepared for launch 0 sec 1 of 1 processes prepared for launch 0 sec releasing clients 0 10 0.00 MB/sec warmup 1 sec latency 5.724 ms 1 77 1.67 MB/sec warmup 2 sec latency 141.678 ms oleg405-client.virtnet: executing wait_import_state_mount (FULL|IDLE) mdc.lustre-MDT0000-mdc-*.mds_server_uuid 1475 0 1 119 1.80 MB/sec warmup 3 sec latency 212.733 ms 1 148 1.56 MB/sec warmup 4 sec latency 197.759 ms mdc.lustre-MDT0000-mdc-*.mds_server_uuid in FULL state after 0 sec 1 181 1.46 MB/sec warmup 5 sec latency 151.646 ms 1 199 1.30 MB/sec warmup 6 sec latency 400.835 ms 1 231 1.26 MB/sec warmup 7 sec latency 193.463 ms 1 271 1.27 MB/sec warmup 8 sec latency 187.700 ms 1 297 1.21 MB/sec warmup 9 sec latency 168.563 ms 1 332 1.20 MB/sec warmup 10 sec latency 196.424 ms 1 365 1.18 MB/sec warmup 11 sec latency 199.575 ms 1 398 1.17 MB/sec warmup 12 sec latency 252.449 ms UUID 1K-blocks Used Available Use% Mounted on lustre-MDT0000_UUID 1414116 2924 1268380 1% /mnt/lustre[MDT:0] lustre-MDT0001_UUID 1414116 2672 1268632 1% /mnt/lustre[MDT:1] lustre-OST0000_UUID 3833116 48872 3553524 2% /mnt/lustre[OST:0] lustre-OST0001_UUID 3833116 8580 3591816 1% /mnt/lustre[OST:1] filesystem_summary: 7666232 57452 7145340 1% /mnt/lustre 1 437 1.18 MB/sec warmup 13 sec latency 150.819 ms 1 477 1.19 MB/sec warmup 14 sec latency 143.747 ms 1 516 1.18 MB/sec warmup 15 sec latency 119.224 ms 1 608 1.31 MB/sec warmup 16 sec latency 85.133 ms 1 669 1.24 MB/sec warmup 17 sec latency 91.609 ms 1 714 1.17 MB/sec warmup 18 sec latency 99.529 ms 1 767 1.12 MB/sec warmup 19 sec latency 91.153 ms test_26 fail mds2 4 times 1 878 0.18 MB/sec execute 1 sec latency 76.740 ms 1 885 0.09 MB/sec execute 2 sec latency 901.540 ms Failing mds2 on oleg405-server 1 914 0.09 MB/sec execute 3 sec latency 175.001 ms Stopping /mnt/lustre-mds2 (opts:) on oleg405-server 1 969 0.08 MB/sec execute 4 sec latency 79.897 ms 1 995 0.09 MB/sec execute 5 sec latency 328.832 ms 1 995 0.07 MB/sec execute 6 sec latency 1331.368 ms 1 995 0.06 MB/sec execute 7 sec latency 2331.570 ms 1 995 0.05 MB/sec execute 8 sec latency 3332.834 ms 1 995 0.05 MB/sec execute 9 sec latency 4333.133 ms 1 995 0.04 MB/sec execute 10 sec latency 5333.324 ms 1 995 0.04 MB/sec execute 11 sec latency 6336.750 ms 03:28:55 (1785223735) shut down facet: mds2 facet_host: oleg405-server facet_failover_host: oleg405-server 1 995 0.04 MB/sec execute 12 sec latency 7336.955 ms 1 995 0.03 MB/sec execute 13 sec latency 8339.822 ms 1 995 0.03 MB/sec execute 14 sec latency 9340.057 ms 1 995 0.03 MB/sec execute 15 sec latency 10343.684 ms 1 995 0.03 MB/sec execute 16 sec latency 11343.945 ms 1 995 0.03 MB/sec execute 17 sec latency 12347.352 ms 1 995 0.02 MB/sec execute 18 sec latency 13347.572 ms 1 995 0.02 MB/sec execute 19 sec latency 14347.778 ms 1 995 0.02 MB/sec execute 20 sec latency 15348.360 ms 1 995 0.02 MB/sec execute 21 sec latency 16348.558 ms Failover mds2 to oleg405-server mount facets: mds2 1 995 0.02 MB/sec execute 22 sec latency 17351.295 ms 1 995 0.02 MB/sec execute 23 sec latency 18351.524 ms 1 995 0.02 MB/sec execute 24 sec latency 19351.860 ms 1 995 0.02 MB/sec execute 25 sec latency 20352.807 ms 1 995 0.02 MB/sec execute 26 sec latency 21355.074 ms 1 995 0.02 MB/sec execute 27 sec latency 22355.274 ms 1 995 0.02 MB/sec execute 28 sec latency 23356.660 ms 1 995 0.01 MB/sec execute 29 sec latency 24357.205 ms Start mds2: mount -t lustre -o localrecov /dev/mapper/mds2_flakey /mnt/lustre-mds2 1 995 0.01 MB/sec execute 30 sec latency 25357.524 ms 1 995 0.01 MB/sec execute 31 sec latency 26357.947 ms 1 995 0.01 MB/sec execute 32 sec latency 27359.629 ms 1 995 0.01 MB/sec execute 33 sec latency 28359.802 ms 1 995 0.01 MB/sec execute 34 sec latency 29360.422 ms oleg405-server: oleg405-server.virtnet: executing set_default_debug -1 all 1 995 0.01 MB/sec execute 35 sec latency 30360.607 ms pdsh@oleg405-client: oleg405-server: ssh exited with exit code 1 1 995 0.01 MB/sec execute 36 sec latency 31360.866 ms Started lustre-MDT0001 03:29:21 (1785223761) targets are mounted 03:29:21 (1785223761) facet_failover done 1 995 0.01 MB/sec execute 37 sec latency 32361.549 ms 1 995 0.01 MB/sec execute 38 sec latency 33361.693 ms 1 997 0.01 MB/sec execute 39 sec latency 34306.441 ms 1 1029 0.04 MB/sec execute 40 sec latency 186.965 ms 1 1083 0.04 MB/sec execute 41 sec latency 79.878 ms 1 1142 0.04 MB/sec execute 42 sec latency 79.132 ms 1 1193 0.04 MB/sec execute 43 sec latency 96.664 ms oleg405-client.virtnet: executing wait_import_state_mount (FULL|IDLE) mdc.lustre-MDT0001-mdc-*.mds_server_uuid 1475 0 1 1233 0.04 MB/sec execute 44 sec latency 75.459 ms mdc.lustre-MDT0001-mdc-*.mds_server_uuid in FULL state after 0 sec 1 1274 0.04 MB/sec execute 45 sec latency 106.739 ms 1 1419 0.10 MB/sec execute 46 sec latency 95.435 ms 1 1467 0.10 MB/sec execute 47 sec latency 196.091 ms 1 1522 0.10 MB/sec execute 48 sec latency 227.099 ms 1 1603 0.10 MB/sec execute 49 sec latency 69.464 ms 1 1656 0.09 MB/sec execute 50 sec latency 90.424 ms 1 1725 0.09 MB/sec execute 51 sec latency 76.885 ms 1 1789 0.09 MB/sec execute 52 sec latency 94.494 ms 1 1851 0.09 MB/sec execute 53 sec latency 78.858 ms 1 1910 0.09 MB/sec execute 54 sec latency 86.848 ms 1 1980 0.09 MB/sec execute 55 sec latency 133.467 ms 1 2016 0.09 MB/sec execute 56 sec latency 120.295 ms 1 2083 0.09 MB/sec execute 57 sec latency 83.175 ms 1 2148 0.09 MB/sec execute 58 sec latency 89.608 ms 1 2178 0.09 MB/sec execute 59 sec latency 107.125 ms 1 2253 0.09 MB/sec execute 60 sec latency 79.706 ms 1 2345 0.10 MB/sec execute 61 sec latency 104.563 ms 1 2386 0.10 MB/sec execute 62 sec latency 81.238 ms 1 2437 0.10 MB/sec execute 63 sec latency 118.905 ms dbench killed by signal 15 stopping dbench on /mnt/lustre at Tue Jul 28 03:29:47 EDT 2026 with return code 0 68432 pts/0 S+ 0:00 dbench -c client.txt 1 -D /mnt/lustre2/d26.replay-dual/run_dbench -t 100 killed dbench main pid 68432 clean dbench files on /mnt/lustre /mnt/lustre /mnt/lustre removed 'client.txt' /mnt/lustre dbench successfully finished