recovery.py June 20, 2014

=== Summary ===
Recovery time:                                   1.41 s
Failure detection time:                          0.33 s
Recovery + detection time:                       1.74 s
Masters:                                            5
Backups:                                           10
Total nodes:                                       10
Replicas:                                           3
Objects per master:                           592950.00 avg /  593336 max
Object size:                                  1057.63 bytes avg / 1057.63 bytes max
Total recovery segment entries:               2966625
Total live object space:                      2990.34 MB
Total recovery segment space w/ overhead:     2998.84 MB
Storage type:                                 disk
Log directory:                                /home/ouster/.netbeans/remote/rcmaster.scs.stanford.edu/ouster2010-Windows-x86_64/C/Users/ouster/Documents/code/ramcloud/logs/20140620115504

=== Coordinator Time ===
Total:                                        1406.70 ms /   99.48%
Starting recovery on backups:                   17.98 ms /    1.27%
Starting recovery on masters:                    0.27 ms /    0.02%
Tablets recovered:                               0.39 ms /    0.03%
Completing recovery on backups:                  0.04 ms /    0.00%
Get table config:                                5.73 ms /    0.41%
Other:                                        1382.33 ms /   97.75%
Receiving in transport:                          1.48 ms /    0.10%

=== Recovery Master Time ===
Total (versus end-to-end recovery time):      1386.63 ms avg / 1388.79 ms max /   98.06% avg
Total:                                        1386.63 ms avg / 1388.79 ms max /  100.00% avg
Waiting for incoming segments:                 316.87 ms avg /  462.98 ms max /   22.85% avg
Inside recoverSegment:                         983.57 ms avg / 1105.80 ms max /   70.93% avg
Final log sync time:                            11.19 ms avg /   12.44 ms max /    0.81% avg
Removing tombstones:                             0.00 ms /    0.00%
Other:                                          75.00 ms avg /   81.74 ms max /    5.30% avg

=== Recovery Master recoverSegment Time ===
Total (versus end-to-end recovery time):       983.57 ms avg / 1105.80 ms max /   69.55% avg
Total:                                         983.57 ms avg / 1105.80 ms max /  100.00% avg
Managing replication:                           32.18 ms avg /   37.50 ms max /    3.27% avg
Verify checksum:                               293.23 ms avg /  335.06 ms max /   29.81% avg
Segment append:                                309.31 ms avg /  337.81 ms max /   31.45% avg
Other:                                         348.85 ms avg /  395.44 ms max /   35.47% avg

=== Recovery Master ReplicaManager Time during recoverSegment ===
Total (versus end-to-end recovery time):        32.18 ms avg /   37.50 ms max /    2.28% avg
Total:                                          32.18 ms avg /   37.50 ms max /  100.00% avg
Posting write RPCs for TX to transport:         12.04 ms avg /   15.29 ms max /   37.40% avg
Other:                                          20.15 ms avg /   22.22 ms max /   62.60% avg

=== Recovery Master Stats ===
Final log sync amount:                          22.83 MB avg /   21.60 MB min /  114.16 MB total
Total replication amount:                     2097.98 MB avg / 2095.33 MB min / 10489.88 MB total
Total replication during replay:              2075.14 MB avg / 2071.97 MB min / 10375.72 MB total
Opening sessions:                                1.09 ms avg /    1.20 ms max /    0.08% avg
Receiving in transport:                          4.48 ms avg /    5.00 ms max /    0.32% avg
Transmitting in transport:                     118.81 ms avg /  125.36 ms max /    8.40% avg
Client RPCs Active:                           1317.01 ms avg / 1393.61 ms max /   93.13% avg
Average GRD completion time:                    11.73 ms avg /   12.11 ms max
Log replication rate:                          581.81 MB/s avg /  566.44 MB/s min / 2909.03 MB/s total
Log replication rate during replay:            580.88 MB/s avg /  565.14 MB/s min / 2904.39 MB/s total
Log replication rate during log sync:          684.16 MB/s avg /  579.00 MB/s min / 3420.81 MB/s total
Replication:                                  1203.00 ms avg / 1234.86 ms max /   85.07% avg
TX active:                                    1091.44 ms avg / 1121.85 ms max /   77.18% avg
TX active during replication:                 1091.19 ms avg / 1121.72 ms max /   90.71% avg
TX active during replay:                      1081.78 ms avg / 1110.13 ms max /   90.77% avg
TX active during log sync:                       9.41 ms avg /   11.60 ms max /   84.06% avg
TX active rate during replication:            1923.54 MB/s avg / 1869.09 MB/s min / 9617.68 MB/s total
TX active rate during replay:                 1919.12 MB/s avg / 1869.16 MB/s min / 9595.60 MB/s total
TX active rate during log sync:               2474.04 MB/s avg / 1862.48 MB/s min / 12370.21 MB/s total
Copying for TX during replication:             115.32 ms avg /  120.74 ms max /    8.15% avg
Copying for TX during replay:                  115.31 ms avg /  120.74 ms max /    8.15% avg
Copying for TX during log sync:                  0.00 ms avg /    0.00 ms max /    0.00% avg
Copying for tx during replication rate:       6068.06 MB/s avg / 5793.17 MB/s min / 30340.29 MB/s total
Copying for TX during replay rate:            6002.13 MB/s avg / 5733.48 MB/s min / 30010.63 MB/s total
Copying for TX during log sync rate:          3338903.52 MB/s avg / 1564744.12 MB/s min / 16694517.61 MB/s total
Max active replication tasks:                       3 tasks /      15 tasks total
Memory read bandwidth used during replay:     4254.26 MB/s avg / 4185.62 MB/s min / 21271.29 MB/s total
Memory write bandwidth used during replay:    3931.24 MB/s avg / 3807.28 MB/s min / 19656.18 MB/s total

=== Backup Time ===
RPC service time:                              376.43 ms avg /  458.28 ms max /   26.62% avg
startReadingData RPC:                            0.34 ms avg /    0.77 ms max /    0.02% avg
write RPC:                                     371.50 ms avg /  450.30 ms max /   26.27% avg
Write copy:                                    369.21 ms avg /  447.65 ms max /   26.11% avg
Other write RPC:                                 2.29 ms avg /    2.82 ms max /    0.16% avg
getRecoveryData RPC:                             3.50 ms avg /   11.75 ms max /    0.25% avg
Other:                                           1.09 ms avg /   16.01 ms max /    0.08% avg
Transmitting in transport:                     108.84 ms avg /  125.36 ms max /    7.70% avg
Filtering segments:                            319.47 ms avg /  419.94 ms max /   22.59% avg
Reading+filtering replicas:                   1317.49 ms avg / 1385.89 ms max /   93.17% avg
Reading replicas from disk:                   1305.35 ms avg / 1372.73 ms max /   92.31% avg
getRecoveryData completions:                   187.50 avg /     190 max
getRecoveryData retry fraction:                  0.28 avg /    0.72 max

=== Efficiency ===
recoverSegment CPU:                              2.62 ms avg
Writing a segment:                               0.73 ms avg
Memory bandwidth (backup copies):                2.41 GB/s avg /    1.87 GB/s min

=== Network Utilization ===
Aggregate:                                     124.25 Gb/s /   49.70%
Master in:                                       7.59 Gb/s avg /    7.16 Gb/s min /   37.95 Gb/s total
Master out:                                     11.59 Gb/s avg /   11.58 Gb/s min /   57.97 Gb/s total
Master out during replication:                  13.64 Gb/s avg /   13.28 Gb/s min /   68.18 Gb/s total
Master out during log sync:                     16.04 Gb/s avg /   13.57 Gb/s min /   80.18 Gb/s total
Backup in:                                       6.63 Gb/s avg /    5.30 Gb/s min /   66.28 Gb/s total
Backup out:                                      6.63 Gb/s avg /    1.64 Gb/s min /   66.28 Gb/s total

=== Disk Utilization ===
Effective bandwidth:                           226.29 MB/s avg /  214.98 MB/s min / 2262.93 MB/s total
Active bandwidth:                              227.81 MB/s avg /  217.54 MB/s min / 2278.05 MB/s total
Active bandwidth reading:                      229.46 MB/s avg /  215.63 MB/s min / 2294.60 MB/s total
Active bandwidth writing:                      231.96 MB/s avg /  163.90 MB/s min / 2319.56 MB/s total
Disk active time:                               99.33 % avg /  102.01 % max
Disk reading time:                              92.31 % avg /   97.07 % max
Disk writing time:                               7.02 % avg /   12.96 % max

=== Backup Events ===
Segments read:                                  37.40 avg /      38 max
Primary segments loaded:                       187.50 avg /     190 max
Secondary segments loaded:                          0

=== Slowest Servers ===
Backup opens, writes:                         192.168.1.104 / 348.8 ms
Stalled reading segs from backups:            192.168.1.104 / 463.0 ms
Reading from disk:                            192.168.1.107 / 215.6 MB/s
Writing to disk:                              192.168.1.108 / 163.9 MB/s