Spectrum Scale: How to set systemd to auto start GPFS daemons after reboot

GPFS comes with a systemd unit file (/usr/lpp/mmfs/lib/systemd/gpfs.service) that comes with the RPMs by default.

systemctl status gpfs.service


[root@IBMPOWER_HOST1 ~]#  mmgetstate -a

 Node number  Node name        GPFS state
-------------------------------------------
       1      IBMPOWER_HOST1        active
       2      IBMPOWER_HOST2        active
       3      IBMPOWER_HOST3        active
[root@IBMPOWER_HOST1 ~]# systemctl status gpfs.service
● gpfs.service - General Parallel File System
   Loaded: loaded (/usr/lib/systemd/system/gpfs.service; disabled; vendor preset: disabled)
   Active: active (running) since Sun 2020-04-19 00:39:51 EDT; 5 days ago
  Process: 65927 ExecStart=/usr/lpp/mmfs/bin/mmremote startSubsys systemd $STARTSUBSYS_ARGS (code=exited, status=0/SUCCESS)
 Main PID: 65959 (runmmfs)
    Tasks: 689
   Memory: 1.0G
   CGroup: /system.slice/gpfs.service
           ├─65959 /usr/lpp/mmfs/bin/mmksh /usr/lpp/mmfs/bin/runmmfs
           └─66297 /usr/lpp/mmfs/bin/mmfsd

Apr 19 00:39:51 IBMPOWER_HOST1 systemd[1]: Started General Parallel File System.
Apr 19 00:39:55 IBMPOWER_HOST1 mmfs[66297]: [N] CCR: failed to connect to node 9.114.75.217:1191 (sock 53 err 79)
Apr 19 00:39:56 IBMPOWER_HOST1 mmfs[66297]: [N] Connecting to 9.114.75.217 IBMPOWER_HOST3 <c0p0>
Apr 19 00:39:56 IBMPOWER_HOST1 mmfs[66297]: [N] This node (9.114.75.216 (IBMPOWER_HOST1)) is now Cluster Manager for SMPICI_gpfs.IBMPOWER_HOST1.
Apr 19 00:39:56 IBMPOWER_HOST1 mmfs[66297]: [N] mmfsd ready
Apr 19 00:39:57 IBMPOWER_HOST1 mmfs[66297]: [N] Node 9.114.75.217 (IBMPOWER_HOST3) appointed as manager for gpfs_fs.
Apr 20 16:08:19 IBMPOWER_HOST1 mmfs[66297]: [N] Node 9.114.75.215 (IBMPOWER_HOST2) lease renewal is overdue. Pinging to check if it is alive
Apr 20 18:23:10 IBMPOWER_HOST1 mmfs[66297]: [N] Node 9.114.75.217 (IBMPOWER_HOST3) lease renewal is overdue. Pinging to check if it is alive
Apr 20 20:25:15 IBMPOWER_HOST1 mmfs[66297]: [N] Node 9.114.75.217 (IBMPOWER_HOST3) lease renewal is overdue. Pinging to check if it is alive
Apr 20 22:46:10 IBMPOWER_HOST1 mmfs[66297]: [N] Node 9.114.75.217 (IBMPOWER_HOST3) lease renewal is overdue. Pinging to check if it is alive
[root@IBMPOWER_HOST1 ~]#


[root@IBMPOWER_HOST1 ~]# systemctl is-active gpfs.service
active
[root@IBMPOWER_HOST1 ~]# systemctl is-enabled gpfs.service
disabled
[root@IBMPOWER_HOST1 ~]#
[root@IBMPOWER_HOST1 ~]# systemctl is-failed gpfs.service
active
[root@IBMPOWER_HOST1 ~]#



This will return active if it is running properly or failed if an error occurred.
If the unit was intentionally stopped, it may return unknown or inactive


---------------------------------

[root@IBMPOWER_HOST1 ~]# rm -rf /etc/systemd/system/multi-user.target.wants/gpfs.service
[root@IBMPOWER_HOST1 ~]# systemctl enable  gpfs.service
Created symlink from /etc/systemd/system/multi-user.target.wants/gpfs.service to /usr/lib/systemd/system/gpfs.service.
[root@IBMPOWER_HOST1 ~]# ls -alsrt /etc/systemd/system/multi-user.target.wants/gpfs.service
0 lrwxrwxrwx 1 root root 36 Apr 24 03:41 /etc/systemd/system/multi-user.target.wants/gpfs.service -> /usr/lib/systemd/system/gpfs.service
[root@IBMPOWER_HOST1 ~]#

This will create a symbolic link from the system’s copy of the service file (usually in /lib/systemd/system or /etc/systemd/system)
into the location on disk where systemd looks for autostart files (usually /etc/systemd/system/some_target.target.wants.


--------------------------------------------
[root@IBMPOWER_HOST1 ~]# systemctl enable  gpfs.service
Created symlink from /etc/systemd/system/multi-user.target.wants/gpfs.service to /usr/lib/systemd/system/gpfs.service.
[root@IBMPOWER_HOST1 ~]#

[root@IBMPOWER_HOST1 ~]# systemctl enable  gpfs.service
Created symlink from /etc/systemd/system/multi-user.target.wants/gpfs.service to /usr/lib/systemd/system/gpfs.service.
[root@IBMPOWER_HOST1 ~]# ls -alsrt /etc/systemd/system/multi-user.target.wants/gpfs.service
0 lrwxrwxrwx 1 root root 36 Apr 24 03:41 /etc/systemd/system/multi-user.target.wants/gpfs.service -> /usr/lib/systemd/system/gpfs.service
[root@IBMPOWER_HOST1 ~]#


Lets check the status

[root@IBMPOWER_HOST1 ~]# systemctl status gpfs.service
● gpfs.service - General Parallel File System
   Loaded: loaded (/usr/lib/systemd/system/gpfs.service; enabled; vendor preset: disabled)
   Active: active (running) since Fri 2020-04-24 03:35:55 EDT; 10min ago
 Main PID: 59025 (runmmfs)
   CGroup: /system.slice/gpfs.service
           ├─59025 /usr/lpp/mmfs/bin/mmksh /usr/lpp/mmfs/bin/runmmfs
           └─59548 /usr/lpp/mmfs/bin/mmfsd

Apr 24 03:35:55 IBMPOWER_HOST1 systemd[1]: Starting General Parallel File System...
Apr 24 03:35:55 IBMPOWER_HOST1 systemd[1]: Can't open PID file /var/mmfs/gen/runmmfsPid (yet?) after start: No such file or directory
Apr 24 03:35:55 IBMPOWER_HOST1 systemd[1]: Started General Parallel File System.
Apr 24 03:36:00 IBMPOWER_HOST1 mmfs[59548]: [N] Connecting to 9.114.75.215 IBMPOWER_HOST2 <c0p2>
Apr 24 03:36:00 IBMPOWER_HOST1 mmfs[59548]: [N] mmfsd ready
Apr 24 03:36:01 IBMPOWER_HOST1 mmfs[59548]: [N] Connecting to 9.114.75.217 IBMPOWER_HOST3 <c0n0>
[root@IBMPOWER_HOST1 ~]#




[root@IBMPOWER_HOST1 ~]#
[root@IBMPOWER_HOST1 ~]#  mmgetstate -a

 Node number  Node name        GPFS state
-------------------------------------------
       1      IBMPOWER_HOST1        active
       2      IBMPOWER_HOST2        active
       3      IBMPOWER_HOST3        active
[root@IBMPOWER_HOST1 ~]#
[root@IBMPOWER_HOST1 ~]# systemctl is-enabled gpfs.service
enabled
[root@IBMPOWER_HOST1 ~]#


How to auto mount the filesystem when GPFS daemon starts - if you have not done at installation 


export PATH=$PATH:/usr/lpp/mmfs/bin

[root@IBMPOWER_HOST1 ~]#  mmgetstate -a

 Node number  Node name        GPFS state
-------------------------------------------
       1      IBMPOWER_HOST1        active
       2      IBMPOWER_HOST2        active
       3      IBMPOWER_HOST3        active
[root@IBMPOWER_HOST1 ~]# mmchfs gpfs_fs -A yes
mmchfs: Propagating the cluster configuration data to all
  affected nodes.  This is an asynchronous process.
[root@IBMPOWER_HOST1 ~]#

-A {yes | no | automount}
    Indicates when the file system is to be mounted:

    yes
        When the GPFS daemon starts.


Verify:

[root@IBMPOWER_HOST1 ~]# mmlsfs gpfs_fs  -A
flag                value                    description
------------------- ------------------------ -----------------------------------
 -A                 yes                      Automatic mount option
[root@IBMPOWER_HOST1 ~]#

How to Auto-load and auto-mount Spectrum scale in simple steps ?

[root@IBMPOWER_HOST1 ~]# mmchconfig autoload=yes
mmchconfig: Command successfully completed
mmchconfig: Propagating the cluster configuration data to all
  affected nodes.  This is an asynchronous process.
[root@IBMPOWER_HOST1 ~]#

 Reference:
https://www.ibm.com/support/knowledgecenter/STXKQY_5.0.4/com.ibm.spectrum.scale.v5r04.doc/bl1adm_mmchfs.htm
https://www.ibm.com/support/knowledgecenter/en/SSCKLT_2.0.0/UG/sec_ug_starting_scale_in_system_file.html