• View
• Edit
• History
• Print

mod_ibmpower

Description

The Power5/6/7 extensions (maximum of 23 metrics) are contained in a separate DSO module (written in C) called mod_ibmpower. If installed, this DSO module is loaded during runtime/startup of gmond.

This DSO contains mostly IBM Power5/6/7 specific metrics that can be used to monitor those system properties.

Platform availability:

• AIX: minimum AIX5L V5.1 ML04 level or higher required
• Linux: SLES 9 or higher required, RHEL 4 Update 3 or higher required

Config file: /etc/ganglia/conf.d/ibmpower.conf

Available metrics

The following 23 additional metrics are available:

1. capped
2. cpu_entitlement
3. cpu_in_lpar
4. cpu_in_machine
5. cpu_in_pool
6. cpu_pool_id
7. cpu_pool_idle
8. cpu_used
9. disk_read
10. disk_write
11. disk_iops
12. fwversion
13. kernel64bit
14. lpar
15. lpar_name
16. lpar_num
17. modelname
18. oslevel
19. serial_num
20. smt
21. splpar
22. weight
23. cpu_in_syspool
    • for Power6/7 only, at least AIX5L V5.3 TL07 level or higher required
    • on a Power5 system: same value as cpu_in_pool

Detailed description

1. capped
   • Type: String value
   • returns "yes" if the system is a POWER5/6 Shared Processor LPAR which is running in capped mode or "no" otherwise
2. cpu_entitlement
   • Type: Float value
   • returns the Capacity Entitlement of the system in units of physical CPUs
3. cpu_in_lpar
   • Type: Integer value
   • returns the number of CPUs the OS sees in the system. In a Power5/6/7 Shared Processor LPAR this returns the number of virtual CPUs. When SMT is enabled this number is doubled.
4. cpu_in_machine
   • Type: Integer value
   • returns the number of physical CPUs in the whole system
5. cpu_in_pool
   • Type: Integer value
   • returns the number of physical CPUs in the Shared Processor Pool
6. cpu_pool_id
   • Type: Integer value
   • returns the shared processor pool ID this LPAR belongs to
7. cpu_pool_idle
   • Type: Float value
   • returns in fractional numbers of physical CPUs how much the Shared Processor Pool is idle
8. cpu_used
   • Type: Float value
   • returns in fractional numbers of physical CPUs how much compute resources this shared processor LPAR has used since the last time this metric was measured
9. disk_read
   • Type: Float value
   • returns in units of kB the total read I/O per second of the system
10. disk_write
    • Type: Float value
    • returns in units of kB the total write I/O per second of the system
11. disk_iops
    • Type: Float value
    • returns the total number of I/O operations (read + write) per second
12. fwversion
    • Type: String value
    • returns the current firmware version of the IBM Power System
13. kernel64bit
    • Type: String value
    • returns "yes" if the running kernel is a 64-bit kernel or "no" otherwise
14. lpar
    • Type: String value
    • returns "yes" if the system is a LPAR or "no" otherwise
15. lpar_name
    • Type: String value
    • returns the name of the LPAR as defined on the Hardware Management Console (HMC) or some reasonable message otherwise
16. lpar_num
    • Type: Integer value
    • returns the partition ID of the LPAR as defined on the Hardware Management Console (HMC) or some reasonable message otherwise
17. modelname
    • Type: String value
    • returns the machine model name
18. oslevel
    • Type: String value
    • returns the version string as provided by the AIX command 'oslevel‘
19. serial_num
    • Type: String value
    • returns the serial number of the system as provided by the AIX command 'uname‘
20. smt
    • Type: String value
    • returns "yes" if SMT is enabled or "no" otherwise
21. splpar
    • Type: String value
    • returns "yes" if the system is running in a shared processor LPAR or "no" otherwise
22. weight
    • Type: Integer value
    • returns the weight of the LPAR running in uncapped mode
23. cpu_in_syspool (on a Power5 system: same value as cpu_in_pool)
    • Type: Integer value
    • returns the number of cores contained in the system shared processor pool

Config file example

modules {
  module {
    name = "ibmpower_module"
    path = "modibmpower.so"
  }
}

collection_group {
  collect_once = yes
  time_threshold = 1200
  metric {
    name = "fwversion"
    title = "Firmware Version"
  }
  metric {
    name = "kernel64bit"
    title = "Kernel 64 bit?"
  }
  metric {
    name = "lpar"
    title = "LPAR Mode?"
  }
  metric {
    name = "lpar_num"
    title = "LPAR Number"
  }
  metric {
    name = "model_name"
    title = "Machine Name"
  }
  metric {
    name = "serial_num"
    title = "System Serial Number"
  }
  metric {
    name = "splpar"
    title = "Shared Processor LPAR?"
  }
}

collection_group {
  collect_every = 180
  time_threshold = 1200
  metric {
    name = "cpu_in_machine"
    title = "Cores in Machine"
  }
  metric {
    name = "lpar_name"
    title = "LPAR Name"
  }
  metric {
    name = "oslevel"
    title = "Output of 'oslevel -s'"
  }
}

collection_group {
  collect_every = 15
  time_threshold = 180
  metric {
    name = "capped"
    title = "Capped Mode?"
  }
  metric {
    name = "cpu_pool_id"
    title = "Shared processor pool ID of this LPAR"
  }
  metric {
    name = "cpu_entitlement"
    title = "CPU Entitlement"
    value_threshold = 0.01
  }
  metric {
    name = "cpu_in_lpar"
    title = "Number of Virtual CPUs in LPAR"
    value_threshold = 1
  }
  metric {
    name = "cpu_in_pool"
    title = "Number of Cores in Pool"
    value_threshold = 1
  }
  metric {
    name = "cpu_in_syspool"
    title = "Number of Cores in System Pool"
    value_threshold = 1
  }
  metric {
    name = "disk_iops"
    title = "Total number I/O operations per second"
    value_threshold = 1.0
  }
  metric {
    name = "disk_read"
    title = "Total Disk Read I/O per second"
    value_threshold = 1.0
  }
  metric {
    name = "disk_write"
    title = "Total Disk Write I/O per second"
    value_threshold = 1.0
  }
  metric {
    name = "smt"
    title = "SMT enabled?"
  }
  metric {
    name = "weight"
    title = "LPAR Weight"
    value_threshold = 1
  }
}

collection_group {
  collect_every = 15
  time_threshold = 60
  metric {
    name = "cpu_pool_idle"
    title = "CPU Pool Idle"
    value_threshold = 0.0001
  }
  metric {
    name = "cpu_used"
    title = "Physical Cores Used"
    value_threshold = 0.0001
  }
}

Change History AIX

• Version 1.4: Feb 09, 2012
  • added new metric cpu_pool_id (cpu_pool_id_func())
• Version 1.3: Apr 27, 2010
  • added sanity check for cpu_pool_idle_func()
  • added new metric fwversion (fwversion_func())
• Version 1.2: Feb 10, 2010
  • added IO ops/sec metric (disk_iops_func())
  • changed metric type from GANGLIA_VALUE_FLOAT to GANGLIA_VALUE_DOUBLE and changed unit to bytes/sec for disk_read_func() and disk_write_func()
  • added model_name metric (model_name_func())
• Version 1.1: Jan 21, 2010
  • improved cpu_used() function
  • fixed defuncts caused by open pipes (popen() without pclose())
  • added checks for possible libperfstat counter resets in
    • cpu_pool_idle_func()
    • cpu_used_func()
    • disk_read_func()
    • disk_write_func()
• Version 1.0: Dec 11, 2008
  • initial version

Change History Linux

• Version 0.4: Feb 09, 2012
  • added new metric cpu_pool_id (cpu_pool_id_func())
• Version 0.3: Apr 27, 2010
  • added sanity check for cpu_pool_idle_func()
  • added new metric fwversion (fwversion_func())
  • fixed cpu_used_func() for systems which have /proc/ppc64/lparcfg and the purr stanza does exist but returns garbage because the CPU does not have a PURR register, e.g., true for PowerPC970 (JS20, JS21)
• Version 0.2: Feb 10, 2010
  • improved cpu_used() function
  • added IO ops/sec metric (disk_iops_func())
  • changed metric type from GANGLIA_VALUE_FLOAT to GANGLIA_VALUE_DOUBLE and changed unit to bytes/sec for disk_read_func() and disk_write_func()
  • added model_name metric (model_name_func())
• Version 0.1: Dec 11, 2008
  • initial release