OS agent release notes

Download the latest OS agent version.
Note that you do not need to upgrade OS agent with every LPAR2RRD server upgrade. Check here if you need any fix or new feature.
  • AIX or VIOS : lpar2rrd-agent-XXX.ppc.rpm (for any AIX or VIOS version)
  • Linux ppc/x86: lpar2rrd-agent-XXX.noarch.rpm (any Linux platform)
  • Linux Debian: lpar2rrd-agent-XXX._all.deb (any Debian based platform)
  • Solaris x86: lpar2rrd-agent-XXX.solaris-i86pc.tar.gz
  • Solaris Sparc: lpar2rrd-agent-XXX.solaris-sparc.tar.gz

Other OS agent resources

Release notes

  • 7.20-1
    • Linux fix: the agent loses VM UUID (/opt/lpar2rrd-agent/.uuid) after upgrade (it cannot map the VM to the OS agent data then in the product UI), it is also not refreshed after reboot
      To avoid it do not use "rpm -Uvh ...", but remove original rpm and install this one like here:
      rpm -e lpar2rrd-agent-7.20-0.noarch; rpm -i lpar2rrd-agent-7.20-1.noarch.rpm
      For future upgrades you can use again rpm -Uvh ...
    • Solaris10: fixed an issue with zone in the command prstat (agent not sending data)
  • 7.20-0
    • Linux: new total IOPS, Data and Latency graphs
    • Linux multipath: using sudo for "multipath" command (only root can run it)
  • 7.00-4
    • AIX: SAN: fcstat is parsing correctly even if there is NVMe stats
    • AIX: fixed SMT number recognization
    • Solaris10: fix for: swap -lk
  • 7.00-1
    • Linux: fix in send of volume_id, not all disks were selected
  • 7.00-0
  • 6.15-2
    • AIX VIOS: unused LAN adapters might produce "SEA LOOP" error log entries, fixed
    • AIX rpm install error fixed "Error in PREUN scriptlet in rpm package lpar2rrd-agent-6.XX..."
    • Solaris CPU queue stats : number of virtual CPUs was wrong
  • 6.15-0
    • AIX/Linux/Solaris: support of CPU queue statistics
  • 6.11-2
    • Solaris: support for new menu structures in the UI
  • 6.10-0
    • Solaris: support of Pools
    • NMON: support of Linux on Power
    • Linux: shmmem fix
  • 6.01-0
    • AIX: fix for filesystem utilization, it has not been reported when NFS filesystems was there
    • Linux: fix for filesystem utilization, problem with long filesystem names
  • 6.00-0
    • AIX WLM (Workload Manager) monitoring
    • Monitoring and alerting for filesystem space utilization
  • 5.05-7
    • AIX pinned memory did not count large pages (an issue since "vmstat -v" implementation for mem data in 4.95-7)
  • 5.05-6
    • AIX WPAR memory was fixed (an issue since "vmstat -v" implementation for mem data in 4.95-7)
  • 5.05-5
    • Linux memory: buffers memory has been added to cached (FS cache)
  • 5.05-2
    • VMware support for RHEL 5.x where UUID is not accessible from /proc
  • 5.05-0
    • VMware support, OS agents running o VM's are mapped directly to VM's in VMware menu tree
    • JOB TOP, CPU and Memory tracking of running processes graphically in the time
  • 5.00-4
    • AIX: skipping fcs interfaces which has scsi_id=0x0 (lsattr -El), it might lead to hanging of fcstat and entstat cmds
  • 5.00
    • Renamed 4.96
  • 4.96
    • Linux: LAN write throughput was wrong, there was the same value as for read
    • AIX: added support for native SCSI adapters : sisioa0
  • 4.95-7
    • Linux: switched to use `cat /proc/meminfo` instead of `free` for memory stats
    • AIX: memory stats switched to use "vmstat -v" instead of "svmon -G" as before
      (svmon counts FS cache in different way)
    • AIX LAN enhanced about packet count in/out
      It is presented in the UI under LAN tab (below usual MB/sec graphs, LPAR2RRD server v4.95-6+)
  • 4.95-4
    • Solaris: fixed a few issues with memory reporting on Sparc HW
    • WPAR: avoids cmds which are not working in WPAR env and just reports errors in the log
  • 4.95-0
    • Solaris support
  • 4.90-0
    • Network statistic ported on CentOS7 & Red Hat 7 (new "ifconfig -a" format)
  • 4.84-3
    • Randomization of agent connecting to the LPAR2RRD could lead to long times, introduced max time (20 mins) for pushing data
  • 4.84-2
    • HMC agent could stop working under some condition since 4.81 agent version, fixed
  • 4.84-1
    • Fixed memory graphs for some Linux X86 like CentOS 6 which had wrong data (negative). Most of Linux X86 have not been affected
  • 4.84-0
    • Linux x86 support
  • 4.81-1
    • 4.80 and 4.81 reported FCS only if there was any disk locally attached what NPIV and tape dedicated FSC are not
  • 4.81
    • OS agent contacts the daemon randomly after 5 minutes to more spread connecting of the daemon in the time
  • 4.80
    • vSCSI adapters support
    • response times support for all vSCSI and FCS adapters are monitored (requires LPAR2RRd server 4.74+)
    • SMT info from LPARs is passed to the server to support SMT based CPU Workload Estimations
    • as prevetion against hanging external shell commands executed from the agent there is limited number of copies every particular external cmd running concurently in the system
    • SEA fix (it did not sent SEA data for some NMON versions which use a bit different key word in the nmon file)
    • introduced variable which preferes hostname before the lpar name in data send to the daemon (usefull for NMON loads from unmanaged full lpar partitions where is not posible to change lpar name)
      LPAR2RRD_HOSTNAME_PREFER=1 /usr/bin/perl /opt/lpar2rrd-agent/lpar2rrd-agent.pl -n /home/nmon_dir <LPAR2RRD server>
  • 4.70
    • fixed a paging issue on AIX where paging of 64kB pages was considered as standard 4kB pages
      (real paging was higher than displayed one)
    • many of small other fixes, upgrade to 4.70 if possible.
  • 4.60
    • it is implemented support for integrated NMON online grapher (aka nmon.lpar2rrd.com) in you local instance
    • couple of small other fixes, you might keep to use 4.50 if it is working fine.
  • 4.50
    • NMON data support
      • It is able to use already generated NMON data as a source (online and batch mode processing is possible) instead of using agent called commands to get statistical data.
      • It might use NMON data and its own data as the source concurrently.
      • It is able to process collected NMON data from many LPARs in one location as a batch job once a day.
    • WPAR support
    • HMC monitoring
    • It posibble to feed via one OS agen more LPAR2RRD server concurently
  • 4.07
    • fixes FCS problems (it avoids non connected fcs adapters which causing problems like long term query and errpt warnings)
  • 4.06
    • it uses internal timestamp to save time of last saved data on the server side, then it sends newer data only
  • 4.05
    • limit of concurrently running OS agents to 10 on single OS (to do not exhaust max number of running processes if anything stuck)
    • possibility to exclude fcstat (SAN stats), the cmd might cause problems
      crontab line:
      * * * * * FCSTAT=/bin/true; export FCSTAT; /usr/bin/perl /opt/lpar2rrd-agent/lpar2rrd-agent.pl >LPAR2RRD-SERVER<   >PORT< > /var/tmp/lpar2rrd-agent.out 2>&1
  • 4.04
    • fix for LPARs (servers) not managed by the HMC, it has not worked before correctly
  • 4.03
    • fix for AME support, it did not work in 4.00, it is necessary to upgrade even LPAR server to 4.03 release at least
    • fix for memory usage data if AMS is being used
  • 4.00
    • New communication protocol (requires 4.00+ LPAR2RRD server)
    • Data collection enhanced about LAN (eth), SAN (fcs), SEA (VIOS only), paging allocated, AME
  • 1.02
    • initial release which supports memory and paging data only