Data gathered by gpfs.snap on Linux

This topic describes the type of data that is always gathered by the gpfs.snap command on the Linux® platform.

Note: The gpfs.snap command does not collect installation toolkit logs. You can collect these logs by using the installer.snap.py script that is located in the same directory as the installation toolkit. For more information, see Logging and debugging for installation toolkit.
These items are always obtained by the gpfs.snap command when gathering data for a Linux node:

  1. The output of these commands:

    • dmesg
    • fdisk -l
    • lsmod
    • lspci
    • rpm -qa
    • rpm --verify gpfs.base
    • rpm --verify gpfs.docs
    • rpm --verify gpfs.gpl
    • rpm --verify gpfs.msg.en_US

  2. The content of these files:

    • /etc/filesystems
    • /etc/fstab
    • /etc/*release
    • /proc/cpuinfo
    • /proc/version
    • /usr/lpp/mmfs/src/config/site.mcr
    • /var/log/messages*

The following data is also collected on Linux on Z:

  1. The output of the dbginfo.sh tool.

    If s390-tools are installed, then the output of dbginfo.sh is captured.

  2. The content of these files:

    • /boot/config-$(active-kernel). For example, /boot/config-3.10.0-123.6.3.el7.s390x

Performance monitoring data

The following data is collected to enable performance monitoring diagnosis:

  1. The output of these commands:

    • mmperfmon config show
    • ps auxw | grep ZIMon
    • service pmsensors status
    • service pmcollector status
    • mmhealth node show perfmon -v
    • du -h /opt/IBM/zimon
    • ls -laR /opt/IBM/zimon/data
    • Start of changemmdiag --waiters --iohist --threads --stats --memoryEnd of change
    • Start of changemmfsadm eventsExporter mmpmon chmsEnd of change
    • Start of changemmfsadm dump nsdEnd of change
    • Start of changemmfsadm dump mbEnd of change

  2. The content of these files:

    • /var/log/zimon/*
    • /opt/IBM/zimon/*.cfg
  3. Start of changeThe outputs of these commands are collected once for the cluster:
    • mmperfmon query --list=keys --raw
    End of change

Call home configuration data

The following data is collected to enable call home diagnosis:

  1. The content of these files is collected for each relevant node:

    • /var/mmfs/tmp/mmcallhome.log
    • /var/mmfs/tmp/callhome/log/callhomeutils.log
    • /var/mmfs/callhome/*

  2. The output of these commands is collected once for the cluster:

    • mmcallhome capability list
    • mmcallhome group list
    • mmcallhome info list
    • mmcallhome proxy list
    • mmcallhome schedule list
    • mmcallhome status list

  3. The output of the mmcallhome test connection command is collected once for each relevant node.

GUI data

The following data is collected to enable performance monitoring diagnosis:

  • The output of these commands:

    • pg_dump -U postgres -h 127.0.0.1 -n fscc postgres
    • /usr/lpp/mmfs/gui/bin/get_version
    • getent passwd scalemgmt
    • getent group scalemgmt
    • iptables -L -n
    • iptables -L -n -t nat
    • systemctl kill gpfsgui --signal=3 --kill-who=main # trigger a core dump
    • systemctl status gpfsgui
    • journalctl _SYSTEMD_UNIT=gpfsgui.service --no-pager -l

  • The content of these files:

    • /etc/sudoers
    • /etc/sysconfig/gpfsgui
    • /opt/ibm/wlp/usr/servers/gpfsgui/*.xml
    • /var/lib/pgsql/data/*.conf
    • /var/lib/pgsql/data/pg_log/*
    • /var/lib/mmfs/gui/*
    • /var/log/cnlog/*
    • /var/crash/scalemgmt/javacore*
    • /var/crash/scalemgmt/heapdump*
    • /var/crash/scalemgmt/Snap*
    • /usr/lpp/mmfs/gui/conf/*

  • The output of these commands is collected once for the cluster:

    • /usr/lpp/mmfs/lib/ftdc/mmlssnap.sh

  • The content of these CCR files is collected once for the cluster:

    • _gui.settings
    • _gui.user.repo
    • _gui.dashboards
    • _gui.snapshots
    • key-value pair: gui_master_node

System monitor data

The following data is collected to help analyze the monitored system data:
  • The output of these commands is collected for each relevant node:
    • mmhealth node eventlog
    • mhealth node show

  • The contents of these files:
    • /var/adm/ras/mmsysmonitor.*.log*
    • /var/adm/ras/top_data*
    • /var/mmfs/tmp/mmhealth.log
    • /var/mmfs/tmp/debugmmhealth.log
    • /var/mmfs/mmsysmon/mmsysmonitor.conf

  • The output of these commands is collected once for the cluster:
    • tsctl shownodes up
    • mmhealth cluster show
    • mmccr flist

  • The contents of the mmsysmon.json CCR file.

InfiniBand interface data

The output of the following commands is collected on Linux nodes with InfiniBand network interface in case of extended network discovery:

  • ibstat
  • iblinkinfo
  • ibdev2netdev
  • ibnetdiscover
  • ip a
  • cat /proc/net/dev
  • ls -l /sys/class/infiniband/

The output of the ibdiagnet command is collected once for the cluster. For more information, see the Data gathered by gpfs.snap for a controller snapshot section.