Please Review: PR 4891 - Issue 4890 - Need cli to easely get simple performance statistics
https://github.com/389ds/389-ds-base/pull/4891
As described in the PR te changes is about adding a CLI and an lib389 Module to have a press button tools to collect some basic performance measure (to get an idea of the bdb versus mdb impact ... )
FYI results with 2 run using ASAN debug build on my labtop are:
*test name* *threads* *users* *bdb* *mdb* *%* *bdb* *mdb* *%* export 100 81.02 61.54 -24.05% 80.43 61.47 -23.58% export 1000 670.22 523.59 -21.88% 661.40 521.54 -21.15% export 10000 3,958.79 3,496.89 -11.67% 3,973.71 3,494.91 -12.05% export 100000 8,338.82 8,612.39 3.28% 8,391.60 8,603.52 2.53% export 1000000 10,310.60 10,877.96 5.50% 10,350.75 10,924.19 5.54% import 100 40.57 36.19 -10.81% 40.37 36.79 -8.85% import 1000 344.93 145.78 -57.74% 344.87 146.36 -57.56% import 10000 1,879.10 237.11 -87.38% 1,876.24 235.01 -87.47% import 100000 3,992.51 1,794.74 -55.05% 4,008.33 1,775.69 -55.70% import 1000000 3,964.03 5,581.28 40.80% 3,986.71 5,586.30 40.12% modify_sn 1 100 234.16 259.77 10.94% 235.29 325.64 38.40% modify_sn 1 1000 233.77 210.26 -10.06% 233.54 194.40 -16.76% modify_sn 1 10000 238.01 142.58 -40.10% 237.82 140.86 -40.77% modify_sn 1 100000 250.42 141.73 -43.41% 250.19 141.31 -43.52% modify_sn 1 1000000 240.70 140.68 -41.56% 245.26 140.26 -42.81% modify_sn 4 100 50.90 25.64 -49.63% 49.38 24.50 -50.38% modify_sn 4 1000 49.55 61.50 24.12% 46.80 93.33 99.41% modify_sn 4 10000 48.20 83.03 72.26% 42.53 80.10 88.36% modify_sn 4 100000 43.38 83.75 93.08% 41.63 81.94 96.86% modify_sn 4 1000000 39.22 81.56 107.93% 40.93 79.25 93.63% modify_sn 8 100 75.64 48.58 -35.77% 73.95 11.31 -84.70% modify_sn 8 1000 73.89 87.35 18.22% 72.00 72.75 1.04% modify_sn 8 10000 73.32 72.58 -1.02% 71.76 91.13 26.98% modify_sn 8 100000 71.64 91.33 27.48% 66.11 90.07 36.24% modify_sn 8 1000000 70.36 90.41 28.49% 67.78 88.92 31.20% search_uid 1 100 6,998.27 6,962.87 -0.51% 7,022.91 6,913.89 -1.55% search_uid 1 1000 6,798.95 3,855.83 -43.29% 6,763.09 3,734.33 -44.78% search_uid 1 10000 6,633.50 3,625.00 -45.35% 6,620.20 3,480.63 -47.42% search_uid 1 100000 6,216.48 3,510.27 -43.53% 6,359.92 3,437.73 -45.95% search_uid 1 1000000 5,997.79 3,251.73 -45.78% 6,055.08 3,212.93 -46.94% search_uid 4 100 8,788.08 8,796.16 0.09% 8,860.81 8,782.21 -0.89% search_uid 4 1000 8,625.73 7,106.80 -17.61% 8,701.26 11,109.74 27.68% search_uid 4 10000 8,561.08 6,831.86 -20.20% 8,827.34 6,943.69 -21.34% search_uid 4 100000 8,255.40 6,728.26 -18.50% 8,484.28 10,849.19 27.87% search_uid 4 1000000 7,771.35 10,035.92 29.14% 7,618.45 10,231.04 34.29% search_uid 8 100 11,854.08 12,010.55 1.32% 11,878.36 11,915.58 0.31% search_uid 8 1000 11,650.58 9,931.17 -14.76% 11,681.74 9,561.99 -18.15% search_uid 8 10000 11,630.03 9,460.23 -18.66% 11,703.93 9,528.57 -18.59% search_uid 8 100000 11,176.31 9,157.01 -18.07% 11,424.31 9,281.65 -18.76% search_uid 8 1000000 10,448.66 8,886.82 -14.95% 10,413.28 8,806.54 -15.43% search_uid_in_dn 1 100 406.79 420.57 3.39% 403.68 420.31 4.12% search_uid_in_dn 1 1000 25.90 2.00 -92.28% 36.89 3.10 -91.60% search_uid_in_dn 1 10000 2.00 79.20 3860.00% 1.60 80.80 4950.00% search_uid_in_dn 1 100000 76.20 #DIV/0! 78.40 #DIV/0! search_uid_in_dn 1 1000000 75.30 #DIV/0! 78.70 #DIV/0! search_uid_in_dn 4 100 1,195.04 1,048.67 -12.25% 1,184.81 1,047.66 -11.58% search_uid_in_dn 4 1000 15.92 2.83 -82.25% 15.50 2.23 -85.65% search_uid_in_dn 4 10000 1.63 42.73 2529.23% 1.75 43.73 2398.57% search_uid_in_dn 4 100000 0.10 41.90 41800.00% 0.13 43.25 34500.00% search_uid_in_dn 4 1000000 0.25 41.75 16600.00% 0.10 42.33 42225.00% search_uid_in_dn 8 100 1,479.11 1,483.01 0.26% 1,475.27 1,486.45 0.76% search_uid_in_dn 8 1000 68.19 6.79 -90.04% 31.73 6.88 -78.33% search_uid_in_dn 8 10000 6.75 19.86 194.24% 4.11 21.81 430.34% search_uid_in_dn 8 100000 0.62 19.61 3063.39% 0.39 20.70 5235.05% search_uid_in_dn 8 1000000 0.14 19.40 13957.97% 0.01 19.50 162400.00%
Looks good!
It may be good to indicate if the difference is bdb - mdb or the other way. AT the moment the -55% doesn't mean much as it could go either way.
Could also be good to alternate a colour when it's pos/negative (light pink works and contrasts well to yellow).
It's also worth noting in these if these are cache hot or cache cold tests, and how we are setting up the entry cache. If anything to really stress bdb/mdb we want the entry cache set to a similar level and as small as possible to push more traffic through these layers,
Otherwise, looks good ! Great stuff :)
On 1 Sep 2021, at 22:38, Pierre Rogier progier@redhat.com wrote:
https://github.com/389ds/389-ds-base/pull/4891
As described in the PR te changes is about adding a CLI and an lib389 Module to have a press button tools to collect some basic performance measure (to get an idea of the bdb versus mdb impact ... )
FYI results with 2 run using ASAN debug build on my labtop are:
test name threads users bdb mdb % bdb mdb % export 100 81.02 61.54 -24.05% 80.43 61.47 -23.58% export 1000 670.22 523.59 -21.88% 661.40 521.54 -21.15% export 10000 3,958.79 3,496.89 -11.67% 3,973.71 3,494.91 -12.05% export 100000 8,338.82 8,612.39 3.28% 8,391.60 8,603.52 2.53% export 1000000 10,310.60 10,877.96 5.50% 10,350.75 10,924.19 5.54% import 100 40.57 36.19 -10.81% 40.37 36.79 -8.85% import 1000 344.93 145.78 -57.74% 344.87 146.36 -57.56% import 10000 1,879.10 237.11 -87.38% 1,876.24 235.01 -87.47% import 100000 3,992.51 1,794.74 -55.05% 4,008.33 1,775.69 -55.70% import 1000000 3,964.03 5,581.28 40.80% 3,986.71 5,586.30 40.12% modify_sn 1 100 234.16 259.77 10.94% 235.29 325.64 38.40% modify_sn 1 1000 233.77 210.26 -10.06% 233.54 194.40 -16.76% modify_sn 1 10000 238.01 142.58 -40.10% 237.82 140.86 -40.77% modify_sn 1 100000 250.42 141.73 -43.41% 250.19 141.31 -43.52% modify_sn 1 1000000 240.70 140.68 -41.56% 245.26 140.26 -42.81% modify_sn 4 100 50.90 25.64 -49.63% 49.38 24.50 -50.38% modify_sn 4 1000 49.55 61.50 24.12% 46.80 93.33 99.41% modify_sn 4 10000 48.20 83.03 72.26% 42.53 80.10 88.36% modify_sn 4 100000 43.38 83.75 93.08% 41.63 81.94 96.86% modify_sn 4 1000000 39.22 81.56 107.93% 40.93 79.25 93.63% modify_sn 8 100 75.64 48.58 -35.77% 73.95 11.31 -84.70% modify_sn 8 1000 73.89 87.35 18.22% 72.00 72.75 1.04% modify_sn 8 10000 73.32 72.58 -1.02% 71.76 91.13 26.98% modify_sn 8 100000 71.64 91.33 27.48% 66.11 90.07 36.24% modify_sn 8 1000000 70.36 90.41 28.49% 67.78 88.92 31.20% search_uid 1 100 6,998.27 6,962.87 -0.51% 7,022.91 6,913.89 -1.55% search_uid 1 1000 6,798.95 3,855.83 -43.29% 6,763.09 3,734.33 -44.78% search_uid 1 10000 6,633.50 3,625.00 -45.35% 6,620.20 3,480.63 -47.42% search_uid 1 100000 6,216.48 3,510.27 -43.53% 6,359.92 3,437.73 -45.95% search_uid 1 1000000 5,997.79 3,251.73 -45.78% 6,055.08 3,212.93 -46.94% search_uid 4 100 8,788.08 8,796.16 0.09% 8,860.81 8,782.21 -0.89% search_uid 4 1000 8,625.73 7,106.80 -17.61% 8,701.26 11,109.74 27.68% search_uid 4 10000 8,561.08 6,831.86 -20.20% 8,827.34 6,943.69 -21.34% search_uid 4 100000 8,255.40 6,728.26 -18.50% 8,484.28 10,849.19 27.87% search_uid 4 1000000 7,771.35 10,035.92 29.14% 7,618.45 10,231.04 34.29% search_uid 8 100 11,854.08 12,010.55 1.32% 11,878.36 11,915.58 0.31% search_uid 8 1000 11,650.58 9,931.17 -14.76% 11,681.74 9,561.99 -18.15% search_uid 8 10000 11,630.03 9,460.23 -18.66% 11,703.93 9,528.57 -18.59% search_uid 8 100000 11,176.31 9,157.01 -18.07% 11,424.31 9,281.65 -18.76% search_uid 8 1000000 10,448.66 8,886.82 -14.95% 10,413.28 8,806.54 -15.43% search_uid_in_dn 1 100 406.79 420.57 3.39% 403.68 420.31 4.12% search_uid_in_dn 1 1000 25.90 2.00 -92.28% 36.89 3.10 -91.60% search_uid_in_dn 1 10000 2.00 79.20 3860.00% 1.60 80.80 4950.00% search_uid_in_dn 1 100000 76.20 #DIV/0! 78.40 #DIV/0! search_uid_in_dn 1 1000000 75.30 #DIV/0! 78.70 #DIV/0! search_uid_in_dn 4 100 1,195.04 1,048.67 -12.25% 1,184.81 1,047.66 -11.58% search_uid_in_dn 4 1000 15.92 2.83 -82.25% 15.50 2.23 -85.65% search_uid_in_dn 4 10000 1.63 42.73 2529.23% 1.75 43.73 2398.57% search_uid_in_dn 4 100000 0.10 41.90 41800.00% 0.13 43.25 34500.00% search_uid_in_dn 4 1000000 0.25 41.75 16600.00% 0.10 42.33 42225.00% search_uid_in_dn 8 100 1,479.11 1,483.01 0.26% 1,475.27 1,486.45 0.76% search_uid_in_dn 8 1000 68.19 6.79 -90.04% 31.73 6.88 -78.33% search_uid_in_dn 8 10000 6.75 19.86 194.24% 4.11 21.81 430.34% search_uid_in_dn 8 100000 0.62 19.61 3063.39% 0.39 20.70 5235.05% search_uid_in_dn 8 1000000 0.14 19.40 13957.97% 0.01 19.50 162400.00%
--
389 Directory Server Development Team _______________________________________________ 389-devel mailing list -- 389-devel@lists.fedoraproject.org To unsubscribe send an email to 389-devel-leave@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/389-devel@lists.fedoraproject.... Do not reply to spam on the list, report it: https://pagure.io/fedora-infrastructure
-- Sincerely,
William Brown
Senior Software Engineer, Identity and Access Management SUSE Labs, Australia
389-devel@lists.fedoraproject.org
-
Pierre Rogier -
William Brown