Extending the performance

Initially have a number of specific data query. In addition to the initial data query, you can expand the query threshold using the PME Pack for XenServer 6.1 which provides a number of additional metrics query.

Among the data to highlight the Supplemental Pack: http://support.citrix.com/article/CTX135033

Installing Supplemental Pack

Copy the contents of the supplemental CD pack in our server XenServer 6.1.

After the copy, we entered into our server via CLI and execute:

Installed, the pack, we have a new tool for analyzing rrdd2csv performance data.

Getting performance data

If you run the command alone, this will show us all obtainable data in CSV format every 5 seconds.

We can see an example of sampling in the next line.

2012-12-18T10: 37:25 Z, 0.016200, 1227525.875000, 0, 3200204800, 0, 0, 0, 16634.853500, 0, 0, 0, 0, 5596.631800, 3802.557400, 0.057600, 46500, 188.268200, 0, 268435456, 268435456, 0, 9250.524400, 0.006900, 1084316, 2147483648, 107.514400, 64.629000, 0, 0, 0, 15684.230500, 0, 0, 2147434496, 0.012200, 0.009600, 393500, 0, 4370.940900, 471.596300, 0, 16687.380900, 0, 0, 1073692672, 0.017300, 225.753200, 41.070700, 0, 536821760, 0, 0, 0, 0, 0.005400, 659488, 0, 1073692672, 0, 0, 0, 0, 0, 0, 188.268200, 0, 0.006100, 283678, 358.887600, 120.532400, 0, 3.397200, 0, 3.397200, 0, 2147434496, 0, 0, 0.057100, 0.030500, 780140544, 774373376, 0.067200, 0.048800, 0.031900, 0.027200, 3392143360, -191889408, 0.073900, 2.213600, 0, 0.085300, 0.242700, 0.242700, 0 , 1.117700, 0.528900, 0.589600, 0.091900, 0, 0, 0, 0, 0, 0, 0, 0, 0.014000, 0.099100, 0, 0, 0, 0, 0, 0, 0.001200, 0, 0.006300, 0.089600, 0 , 0, 0, 0, 0, 0, 0, 0, 0.011200, 0.090900, 0, 0, 0, 0, 0, 0, 0, 0, 0.002400, 0.806100, 0.027900, 0.048200, 0.116600, 0.808300, 0.025500, 0.048200 , 0.118100, 0.808300, 0.029000, 0.049200, 0.113500, 0.819600, 0.025000, 0.051800, 0.104400, 0, 0, 0, 0.006100, 0.094300, 0.089100, 0.102200, 0.108300, 0.098500, 0, 0, 3870.226600, 1064.562900, 3419.961400, 3419.961400, 7290.188000 , 4484.524400, 1175.309300, 412, 579, 1294.094200, 5410440, 16750624

As with any Linux, the command display help rrdd2csv us online or manually, by typing man command.

Here as well, some of the data we can draw.

SYNOPSIS

      rrd2csv [options] [-datasource descriptor] *

DESCRIPTION

       rrd2csv is a tool to output values ​​XenCloudPlatform and XenServer RRD

       in CSV format. It runs inside dom0, and by default displays all active

       Royal Decrees That Maintains datasources, but it is possible to select Also

       datasources of interest as command-line arguments.

       Datasources (or DS) are DRR concepts. They are objects That have mea-

       surements Associated with them. They include a metric, a source from

       Which the metric originate, and a consolidation function. For

       instance, if You have multiple VMs, you Will Have as many «memory_tar-

       get «You have datasources as VMs.

OPTIONS

       -N Show name labels instead of UUIDs

       -S interval SECONDS

       -U Show UUIDs as well as name labels.

DATASOURCE DESCRIPTOR

              DS :: = CF «,» source «:» UUID «:» METRIC

              CF :: = «AVERAGE» | «MIN» | «MAX» | «LAST»

              SOURCE :: = «host» | «vm»

              UUID :: = string, e.g. «6266c08b-1992-696e-80ed-68ebdd4c36f4»

              METRIC :: = string, e.g. «MEMORY_TARGET»

       UUID is the UUID of the XenAPI’s source and the metric METRIC name. If

       you omit any of the fields, This Will act as a wildcard. As an exam-

       ple, Specifying «:::» will select all enabled datasources, Whereas

       «Vm ::» will select only the datasources related to VMs. Multiple

       datasources can be specified on the command line, the result will be

       the union of single filter results.

Examples man:

       Select host related metrics:

              $ Rrd2csv: host ::

       Select VM related metrics:

              $ Rrd2csv: vm ::

       Select host Metrics that use AVERAGE function as consolidation and any

       That metric has name «MEMORY_TARGET» as consolidation and MIN function:

              $ Rrd2csv AVERAGE: host :: MIN ::: MEMORY_TARGET

Initially we get the list of parameters available in our metric generator to view and select those counters available. Both host level and at the level of VM.

If we want to get the Average memory of all VMs including host, use:

rrd2csv MEMORY_TARGET ::: AVERAGE

If we want that instead of the name appears UUID of the VM use:

rrd2csv -n ::: AVERAGE MEMORY_TARGET

Analysis overall CPU

We will generate a small overall analysis of CPU (CPUs all VMs) through this command. To do this, execute:

 rrd2csv -n -s cpu ::: AVERAGE 15 >> cpus_xenserver.csv

 To indicate only «cpu» is fill in the rest of the fields as if we had used a wildcard *, so we will have all the cpus for all VMs in the environment. With-n see the name of your VM UUID instead. And using the-S define the data collection cycle.