[lustre-discuss] Slow mount on clients

Andrew Elwell andrew.elwell at gmail.com
Tue Feb 4 03:14:49 PST 2020


> HA / MGS running on second node in fstab
:-) that was one of the first things we checked, and I've tried
manually mounting it but no change

10.10.36.224 at o2ib4:10.10.36.225 at o2ib4:/askapfs1  3.7P  3.0P  507T  86%
/askapbuffer

hpc-admin2:~ # lctl ping 10.10.36.224 at o2ib4
12345-0 at lo
12345-10.10.36.224 at o2ib4
hpc-admin2:~ # lctl ping 10.10.36.225 at o2ib4
12345-0 at lo
12345-10.10.36.225 at o2ib4
hpc-admin2:~ # umount /askapbuffer
hpc-admin2:~ # time mount /askapbuffer/

real 1m15.099s
user 0m0.012s
sys 0m0.021s
hpc-admin2:~ #

and on the server:

[root at askap-fs1-mds01 ~]# mount -t lustre
/dev/mapper/array00_2 on /lustre/MGS type lustre (ro)
/dev/mapper/array00_1 on /lustre/askapfs1-MDT0001 type lustre (ro)
[root at askap-fs1-mds01 ~]# lctl list_nids
10.10.36.224 at o2ib4
[root at askap-fs1-mds01 ~]# tunefs.lustre --dryrun /dev/mapper/array00_2
checking for existing Lustre data: found
Reading CONFIGS/mountdata

   Read previous values:
Target:     MGS
Index:      unassigned
Lustre FS:  askapfs1
Mount type: ldiskfs
Flags:      0x1004
              (MGS no_primnode )
Persistent mount opts: user_xattr,errors=remount-ro
Parameters: failover.node=10.10.36.224 at o2ib4:10.10.36.225 at o2ib4


   Permanent disk data:
Target:     MGS
Index:      unassigned
Lustre FS:  askapfs1
Mount type: ldiskfs
Flags:      0x1004
              (MGS no_primnode )
Persistent mount opts: user_xattr,errors=remount-ro
Parameters: failover.node=10.10.36.224 at o2ib4:10.10.36.225 at o2ib4

exiting before disk write.
[root at askap-fs1-mds01 ~]#


(MDT0000 is mounted on the other node at this time).


More information about the lustre-discuss mailing list