iostat2pcp — import iostat data and create a PCP archive
iostat2pcp [-v] [-S start] [-t interval] [-Z timezone] infile outfile
iostat2pcp reads a text file created with iostat(1) (infile) and translates this into a Performance Co-Pilot (PCP) archive with the basename outfile. If infile is “-” then iostat2pcp reads from standard input, allowing easy preprocessing of the iostat(1) output with sed(1) or similar.
The resultant PCP archive may be used with all the PCP client tools to graph subsets of the data using pmchart(1), perform data reduction and reporting, filter with the PCP inference engine pmie(1), etc.
A series of physical files will be created with the prefix outfile. These are outfile.0 (the performance data), outfile.meta (the metadata that describes the performance data) and outfile.index (a temporal index to improve efficiency of replay operations for the archive). If any of these files exists already, then iostat2pcp will not overwrite them and will exit with an error message.
The first output sample from iostat(1) contains a statistical summary since boot time and is ignored by iostat2pcp, so the first real data set is the second one in the iostat(1) output.
The best results are obtained when iostat(1) was run with its own -t flag, so each output sample is prefixed with a timestamp. Even better is -t with $S_TIME_FORMAT=ISO set in environment when iostat(1) is run, in which case the timestamp includes the timezone.
Note that if $S_TIME_FORMAT=ISO is not used with the -t option then iostat(1) may produce a timestamp controlled by LC_TIME from the locale that is in a format iostat2pcp cannot parse. The formats for the timestamp that iostat2pcp accepts are illustrated by these examples:
(for the $S_TIME_FORMAT=ISO).
- 2013-07-06 21:34:39
(for some of the European formats, e.g. de_AT, de_BE, de_LU and en_DK.utf8).
- 06/07/13 21:34:39
(for all of the $LC_TIME settings for English locales outside North America, e.g. en_AU, en_GB, en_IE, en_NZ, en_SG and en_ZA, and all the Spanish locales, e.g. es_ES, es_MX and es_AR).
In particular, note that some common North American $LC_TIME settings will not work with iostat2pcp (namely, en_US, POSIX and C) because they use the MM/DD format which may be incorrectly converted with the assumed DD/MM format. This is another reason to recommend setting $S_TIME_FORMAT=ISO.
If there are no timestamps in the input stream, iostat2pcp will try and deduce the sample interval if basic Disk data (-d option for iostat(1)) is found. If this fails, then the -t option may be used to specify the sample interval in seconds. This option is ignored if timestamps are found in the input stream.
The -S option may be used to specify as start time for the first real sample in infile, where start must have the format HH:MM:SS. This option is ignored if timestamps are found in the input stream.
The -Z option may be used to specify a timezone. It must have the format +HHMM (for hours and minutes East of UTC) or -HHMM (for hours and minutes West of UTC). Note in particular that neither the zoneinfo (aka Olson) format, e.g. Europe/Paris, nor the Posix TZ format, e.g. EST+5 is allowed for the -Z option. This option is ignored if ISO timestamps are found in the input stream. If the timezone is not specified and cannot be deduced, it defaults to “UTC”.
Some additional diagnostic output is generated with the -v option.
iostat2pcp is a Perl script that uses the PCP::LogImport Perl wrapper around the PCP libpcp_import library, and as such could be used as an example to develop new tools to import other types of performance data and create PCP archives.
The available command line options are:
- -S start
Specify the start time for the first sample.
- -t interval
Specify the sample interval in seconds.
Print verbose output.
- -Z timezone
Specify the timezone to use, see above.
iostat2pcp requires infile to have been created by the version of iostat(1) from <http://freshmeat.net/projects/sysstat>.
iostat2pcp handles the -c (CPU), -d (Disk), -x (eXtended Disk) and -p (Partition) report formats (including their -k, -m, -z and ALL variants), but does not accommodate the -n (Network Filesystem) report format from iostat(1); this is a demand-driven limitation rather than a technical limitation.
Environment variables with the prefix PCP_ are used to parameterize the file and directory names used by PCP. On each installation, the file /etc/pcp.conf contains the local values for these variables. The $PCP_CONF variable may be used to specify an alternative configuration file, as described in pcp.conf(5).
iostat(1), pmchart(1), pmie(1), pmlogger(1), sed(1), Date::Format(3pm), Date::Parse(3pm), PCP::LogImport(3pm) and LOGIMPORT(3).