Initialization Options

pgbench accepts the following command-line initialization arguments:

Benchmarking Options

pgbench accepts the following command-line benchmarking arguments:

Common Options

pgbench accepts the following command-line common arguments:

What is the “Transaction” Actually Performed in pgbench?

The default transaction script issues seven commands per transaction:

If you specify -N, steps 4 and 5 aren't included in the transaction. If you specify -S, only the SELECT is issued.

Custom Scripts

pgbench has support for running custom benchmark scenarios by replacing the default transaction script (described above) with a transaction script read from a file (-f option). In this case a “transaction” counts as one execution of a script file. You can even specify multiple scripts (multiple -f options), in which case a random one of the scripts is chosen each time a client session starts a new transaction.

The format of a script file is one SQL command per line; multiline SQL commands are not supported. Empty lines and lines beginning with -- are ignored. Script file lines can also be “meta commands”, which are interpreted by pgbench itself, as described below.

There is a simple variable-substitution facility for script files. Variables can be set by the command-line -D option, explained above, or by the meta commands explained below. In addition to any variables preset by -D command-line options, there are a few variables that are preset automatically, listed in Table 223. A value specified for these variables using -D takes precedence over the automatic presets. Once set, a variable's value can be inserted into a SQL command by writing :variablename. When running more than one client session, each session has its own set of variables.

Script file meta commands begin with a backslash (\). Arguments to a meta command are separated by white space. These meta commands are supported:

As an example, the full definition of the built-in TPC-B-like transaction is:

\set nbranches :scale
\set ntellers 10 * :scale
\set naccounts 100000 * :scale
\setrandom aid 1 :naccounts
\setrandom bid 1 :nbranches
\setrandom tid 1 :ntellers
\setrandom delta -5000 5000
BEGIN;
UPDATE pgbench_accounts SET abalance = abalance + :delta WHERE aid = :aid;
SELECT abalance FROM pgbench_accounts WHERE aid = :aid;
UPDATE pgbench_tellers SET tbalance = tbalance + :delta WHERE tid = :tid;
UPDATE pgbench_branches SET bbalance = bbalance + :delta WHERE bid = :bid;
INSERT INTO pgbench_history (tid, bid, aid, delta, mtime) VALUES (:tid, :bid, :aid, :delta, CURRENT_TIMESTAMP);
END;

This script allows each iteration of the transaction to reference different, randomly-chosen rows. (This example also shows why it's important for each client session to have its own variables — otherwise they'd not be independently touching different rows.)

Per-Transaction Logging

With the -l option but without the --aggregate-interval, pgbench writes the time taken by each transaction to a log file. The log file will be named pgbench_log.nnn, where nnn is the PID of the pgbench process. If the -j option is 2 or higher, creating multiple worker threads, each will have its own log file. The first worker will use the same name for its log file as in the standard single worker case. The additional log files for the other workers will be named pgbench_log.nnn.mmm, where mmm is a sequential number for each worker starting with 1.

The format of the log is:

client_id transaction_no time file_no time_epoch time_us [schedule_lag]

where time is the total elapsed transaction time in microseconds, file_no identifies which script file was used (useful when multiple scripts were specified with -f), and time_epoch/time_us are a Unix epoch format time stamp and an offset in microseconds (suitable for creating an ISO 8601 time stamp with fractional seconds) showing when the transaction completed. Field schedule_lag is the difference between the transaction's scheduled start time, and the time it actually started, in microseconds. It is only present when the --rate option is used. When both --rate and --latency-limit are used, the time for a skipped transaction will be reported as skipped.

Here is a snippet of the log file generated:

 0 199 2241 0 1175850568 995598
 0 200 2465 0 1175850568 998079
 0 201 2513 0 1175850569 608
 0 202 2038 0 1175850569 2663

Another example with --rate=100 and --latency-limit=5 (note the additional schedule_lag column):

 0 81 4621 0 1412881037 912698 3005
 0 82 6173 0 1412881037 914578 4304
 0 83 skipped 0 1412881037 914578 5217
 0 83 skipped 0 1412881037 914578 5099
 0 83 4722 0 1412881037 916203 3108
 0 84 4142 0 1412881037 918023 2333
 0 85 2465 0 1412881037 919759 740

In this example, transaction 82 was late, because its latency (6.173 ms) was over the 5 ms limit. The next two transactions were skipped, because they were already late before they were even started.

When running a long test on hardware that can handle a lot of transactions, the log files can become very large. The --sampling-rate option can be used to log only a random sample of transactions.

Aggregated Logging

With the --aggregate-interval option, the logs use a bit different format:

interval_start num_of_transactions latency_sum latency_2_sum min_latency max_latency [lag_sum lag_2_sum min_lag max_lag [skipped_transactions]]

where interval_start is the start of the interval (Unix epoch format time stamp), num_of_transactions is the number of transactions within the interval, latency_sum is a sum of latencies (so you can compute average latency easily). The following two fields are useful for variance estimation - latency_sum is a sum of latencies and latency_2_sum is a sum of 2nd powers of latencies. The next two fields are min_latency - a minimum latency within the interval, and max_latency - maximum latency within the interval. A transaction is counted into the interval when it was committed. The fields in the end, lag_sum, lag_2_sum, min_lag, and max_lag, are only present if the --rate option is used. The very last one, skipped_transactions, is only present if the option --latency-limit is present, too. They are calculated from the time each transaction had to wait for the previous one to finish, i.e., the difference between each transaction's scheduled start time and the time it actually started.

Here is example output:

1345828501 5601 1542744 483552416 61 2573
1345828503 7884 1979812 565806736 60 1479
1345828505 7208 1979422 567277552 59 1391
1345828507 7685 1980268 569784714 60 1398
1345828509 7073 1979779 573489941 236 1411

Notice that while the plain (unaggregated) log file contains a reference to the custom script files, the aggregated log does not. Therefore if you need per script data, you need to aggregate the data on your own.

Per-Statement Latencies

With the -r option, pgbench collects the elapsed transaction time of each statement executed by every client. It then reports an average of those values, referred to as the latency for each statement, after the benchmark has finished.

For the default script, the output will look similar to this:

starting vacuum...end.
transaction type: TPC-B (sort of)
scaling factor: 1
query mode: simple
number of clients: 10
number of threads: 1
number of transactions per client: 1000
number of transactions actually processed: 10000/10000
tps = 618.764555 (including connections establishing)
tps = 622.977698 (excluding connections establishing)
statement latencies in milliseconds:
        0.004386        \set nbranches 1 * :scale
        0.001343        \set ntellers 10 * :scale
        0.001212        \set naccounts 100000 * :scale
        0.001310        \setrandom aid 1 :naccounts
        0.001073        \setrandom bid 1 :nbranches
        0.001005        \setrandom tid 1 :ntellers
        0.001078        \setrandom delta -5000 5000
        0.326152        BEGIN;
        0.603376        UPDATE pgbench_accounts SET abalance = abalance + :delta WHERE aid = :aid;
        0.454643        SELECT abalance FROM pgbench_accounts WHERE aid = :aid;
        5.528491        UPDATE pgbench_tellers SET tbalance = tbalance + :delta WHERE tid = :tid;
        7.335435        UPDATE pgbench_branches SET bbalance = bbalance + :delta WHERE bid = :bid;
        0.371851        INSERT INTO pgbench_history (tid, bid, aid, delta, mtime) VALUES (:tid, :bid, :aid, :delta, CURRENT_TIMESTAMP);
        1.212976        END;

If multiple script files are specified, the averages are reported separately for each script file.

Note that collecting the additional timing information needed for per-statement latency computation adds some overhead. This will slow average execution speed and lower the computed TPS. The amount of slowdown varies significantly depending on platform and hardware. Comparing average TPS values with and without latency reporting enabled is a good way to measure if the timing overhead is significant.

Good Practices

It is very easy to use pgbench to produce completely meaningless numbers. Here are some guidelines to help you get useful results.

In the first place, never believe any test that runs for only a few seconds. Use the -t or -T option to make the run last at least a few minutes, so as to average out noise. In some cases you could need hours to get numbers that are reproducible. It's a good idea to try the test run a few times, to find out if your numbers are reproducible or not.

For the default TPC-B-like test scenario, the initialization scale factor (-s) should be at least as large as the largest number of clients you intend to test (-c); else you'll mostly be measuring update contention. There are only -s rows in the pgbench_branches table, and every transaction wants to update one of them, so -c values in excess of -s will undoubtedly result in lots of transactions blocked waiting for other transactions.

The default test scenario is also quite sensitive to how long it's been since the tables were initialized: accumulation of dead rows and dead space in the tables changes the results. To understand the results you must keep track of the total number of updates and when vacuuming happens. If autovacuum is enabled it can result in unpredictable changes in measured performance.

A limitation of pgbench is that it can itself become the bottleneck when trying to test a large number of client sessions. This can be alleviated by running pgbench on a different machine from the database server, although low network latency will be essential. It might even be useful to run several pgbench instances concurrently, on several client machines, against the same database server.

Security

If untrusted users have access to a database that has not adopted a secure schema usage pattern, do not run pgbench in that database. pgbench uses unqualified names and does not manipulate the search path.