d/mutilate
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
1e78fc3847
mutilate/cmdline.ggo

112 lines
4.8 KiB
Plaintext
Raw Normal View History

README update. (Round 2)
2014-01-15 01:34:49 +00:00
package "mutilate"
Initial check-in. This is extracted from another git repo. This is the first release, and the prior commit history is not terribly interesting, so I'm not going to bother using filter-branch to try to cleanly isolate the history for this tool. Cheers.
2012-08-23 21:30:32 +00:00
version "0.1"
usage "mutilate -s server[:port] [options]"
description "\"High-performance\" memcached benchmarking tool"
args "-c cc --show-required -C --default-optional -l"
option "verbose" v "Verbosity. Repeat for more verbose." multiple
option "quiet" - "Disable log messages."
text "\nBasic options:"
option "server" s "Memcached server hostname[:port]. \
Repeat to specify multiple servers." string multiple
Added binary support, toggable via --binary.
2013-03-01 13:36:35 +00:00
option "binary" - "Use binary memcached protocol instead of ASCII."
Added location parameter to GPareto distribution. Validated value-size distribution against Atikoglu's paper by comparing CDF and PDF.
2012-08-24 21:02:57 +00:00
option "qps" q "Target aggregate QPS. 0 = peak QPS." int default="0"
Initial check-in. This is extracted from another git repo. This is the first release, and the prior commit history is not terribly interesting, so I'm not going to bother using filter-branch to try to cleanly isolate the history for this tool. Cheers.
2012-08-23 21:30:32 +00:00
option "time" t "Maximum time to run (seconds)." int default="5"
option "keysize" K "Length of memcached keys (distribution)."
string default="30"
option "valuesize" V "Length of memcached values (distribution)."
string default="200"
option "records" r "Number of memcached records to use. \
If multiple memcached servers are given, this number is divided \
by the number of servers." int default="10000"
option "update" u "Ratio of set:get commands." float default="0.0"
text "\nAdvanced options:"
Merged Stephen Yang's and David Terei's binary protocol and SASL implementations. I ended up cherry-picking different parts from each implementation. I liked David's changes to the read state machine to implement SASL. I used Stephene's implementation of the binary protocol integrated with the ASCII protocol.
2013-03-04 17:31:13 +00:00
option "username" U "Username to use for SASL authentication." string
option "password" P "Password to use for SASL authentication." string
Initial check-in. This is extracted from another git repo. This is the first release, and the prior commit history is not terribly interesting, so I'm not going to bother using filter-branch to try to cleanly isolate the history for this tool. Cheers.
2012-08-23 21:30:32 +00:00
option "threads" T "Number of threads to spawn." int default="1"
Misc. experimental features. (1) thread affinity (2) event_base_flag_precise_timer (3) configurable agent port (4) per-connection qps moderation
2013-04-15 20:27:06 +00:00
option "affinity" - "Set CPU affinity for threads, round-robin"
Initial check-in. This is extracted from another git repo. This is the first release, and the prior commit history is not terribly interesting, so I'm not going to bother using filter-branch to try to cleanly isolate the history for this tool. Cheers.
2012-08-23 21:30:32 +00:00
option "connections" c "Connections to establish per server." int default="1"
option "depth" d "Maximum depth to pipeline requests." int default="1"
option "roundrobin" R "Assign threads to servers in round-robin fashion. \
By default, each thread connects to every server."
Added location parameter to GPareto distribution. Validated value-size distribution against Atikoglu's paper by comparing CDF and PDF.
2012-08-24 21:02:57 +00:00
option "iadist" i "Inter-arrival distribution (distribution). Note: \
The distribution will automatically be adjusted to match the QPS given \
by --qps." string default="exponential"
Add --skip option: By default, mutilate tries to maintain a long-term QPS average by issuing requests as quickly as possible until it catches up with where it 'ought' to be. If you give the --skip option, it will instead just skip sending requests that are already late.
2013-03-12 22:27:29 +00:00
option "skip" S "Skip transmissions if previous requests are late. This \
harms the long-term QPS average, but reduces spikes in QPS after \
long latency requests."
Misc. experimental features. (1) thread affinity (2) event_base_flag_precise_timer (3) configurable agent port (4) per-connection qps moderation
2013-04-15 20:27:06 +00:00
option "moderate" - "Enforce a minimum delay of ~1/lambda between requests."
Initial check-in. This is extracted from another git repo. This is the first release, and the prior commit history is not terribly interesting, so I'm not going to bother using filter-branch to try to cleanly isolate the history for this tool. Cheers.
2012-08-23 21:30:32 +00:00
option "noload" - "Skip database loading."
option "loadonly" - "Load database and then exit."
option "blocking" B "Use blocking epoll(). May increase latency."
Add --skip option: By default, mutilate tries to maintain a long-term QPS average by issuing requests as quickly as possible until it catches up with where it 'ought' to be. If you give the --skip option, it will instead just skip sending requests that are already late.
2013-03-12 22:27:29 +00:00
option "no_nodelay" - "Don't use TCP_NODELAY."
Initial check-in. This is extracted from another git repo. This is the first release, and the prior commit history is not terribly interesting, so I'm not going to bother using filter-branch to try to cleanly isolate the history for this tool. Cheers.
2012-08-23 21:30:32 +00:00
option "warmup" w "Warmup time before starting measurement." int
option "wait" W "Time to wait after startup to start measurement." int
Option to save latency measurements to a file.
2013-07-01 17:08:16 +00:00
option "save" - "Record latency samples to given file." string
Initial check-in. This is extracted from another git repo. This is the first release, and the prior commit history is not terribly interesting, so I'm not going to bother using filter-branch to try to cleanly isolate the history for this tool. Cheers.
2012-08-23 21:30:32 +00:00
Add --skip option: By default, mutilate tries to maintain a long-term QPS average by issuing requests as quickly as possible until it catches up with where it 'ought' to be. If you give the --skip option, it will instead just skip sending requests that are already late.
2013-03-12 22:27:29 +00:00
option "search" - "Search for the QPS where N-order statistic < Xus. \
Initial check-in. This is extracted from another git repo. This is the first release, and the prior commit history is not terribly interesting, so I'm not going to bother using filter-branch to try to cleanly isolate the history for this tool. Cheers.
2012-08-23 21:30:32 +00:00
(i.e. --search 95:1000 means find the QPS where 95% of requests are \
faster than 1000us)." string typestr="N:X"
option "scan" - "Scan latency across QPS rates from min to max."
string typestr="min:max:step"
text "\nAgent-mode options:"
option "agentmode" A "Run client in agent mode."
option "agent" a "Enlist remote agent." string typestr="host" multiple
Misc. experimental features. (1) thread affinity (2) event_base_flag_precise_timer (3) configurable agent port (4) per-connection qps moderation
2013-04-15 20:27:06 +00:00
option "agent_port" p "Agent port." string default="5556"
1. Implemented --measure_* options to reasonably separate configuration of measurement client (mutilate master) from load generating clients (mutilate agents). Use -C and --measure_depth to approximate open-loop measurement control (i.e. so long latency requests don't screw up the sample population), and use -Q to ensure that you see minimal client-side queuing delay. 2. Quickly documented agent "protocol". Lots of improvement to be had here...
2013-03-08 21:19:47 +00:00
option "lambda_mul" l "Lambda multiplier. Increases share of \
QPS for this client." int default="1"
option "measure_connections" C "Master client connections per server, \
overrides --connections." int
option "measure_qps" Q "Explicitly set master client QPS, \
spread across threads and connections." int
Add --skip option: By default, mutilate tries to maintain a long-term QPS average by issuing requests as quickly as possible until it catches up with where it 'ought' to be. If you give the --skip option, it will instead just skip sending requests that are already late.
2013-03-12 22:27:29 +00:00
option "measure_depth" D "Set master client connection depth." int
1. Implemented --measure_* options to reasonably separate configuration of measurement client (mutilate master) from load generating clients (mutilate agents). Use -C and --measure_depth to approximate open-loop measurement control (i.e. so long latency requests don't screw up the sample population), and use -Q to ensure that you see minimal client-side queuing delay. 2. Quickly documented agent "protocol". Lots of improvement to be had here...
2013-03-08 21:19:47 +00:00
text "
The --measure_* options aid in taking latency measurements of the
memcached server without incurring significant client-side queuing
delay. --measure_connections allows the master to override the
--connections option. --measure_depth allows the master to operate as
an \"open-loop\" client while other agents continue as a regular
closed-loop clients. --measure_qps lets you modulate the QPS the
master queries at independent of other clients. This theoretically
normalizes the baseline queuing delay you expect to see across a wide
range of --qps values.
"
Initial check-in. This is extracted from another git repo. This is the first release, and the prior commit history is not terribly interesting, so I'm not going to bother using filter-branch to try to cleanly isolate the history for this tool. Cheers.
2012-08-23 21:30:32 +00:00
text "
Some options take a 'distribution' as an argument.
Distributions are specified by <distribution>[:<param1>[,...]].
Parameters are not required. The following distributions are supported:
Added location parameter to GPareto distribution. Validated value-size distribution against Atikoglu's paper by comparing CDF and PDF.
2012-08-24 21:02:57 +00:00
[fixed:]<value> Always generates <value>.
uniform:<max> Uniform distribution between 0 and <max>.
normal:<mean>,<sd> Normal distribution.
exponential:<lambda> Exponential distribution.
pareto:<loc>,<scale>,<shape> Generalized Pareto distribution.
gev:<loc>,<scale>,<shape> Generalized Extreme Value distribution.
Initial check-in. This is extracted from another git repo. This is the first release, and the prior commit history is not terribly interesting, so I'm not going to bother using filter-branch to try to cleanly isolate the history for this tool. Cheers.
2012-08-23 21:30:32 +00:00
Added location parameter to GPareto distribution. Validated value-size distribution against Atikoglu's paper by comparing CDF and PDF.
2012-08-24 21:02:57 +00:00
To recreate the Facebook \"ETC\" request stream from [1], the
following hard-coded distributions are also provided:
fb_value = a hard-coded discrete and GPareto PDF of value sizes
fb_key = \"gev:30.7984,8.20449,0.078688\", key-size distribution
fb_ia = \"pareto:0.0,16.0292,0.154971\", inter-arrival time dist.
Initial check-in. This is extracted from another git repo. This is the first release, and the prior commit history is not terribly interesting, so I'm not going to bother using filter-branch to try to cleanly isolate the history for this tool. Cheers.
2012-08-23 21:30:32 +00:00
[1] Berk Atikoglu et al., Workload Analysis of a Large-Scale Key-Value Store,
SIGMETRICS 2012
"
Reference in New Issue Copy Permalink