[LU-8338] lov.*-clilov-*.stripe* params not consistent with FS defaults

• Status: Open
• Project: Lustre
• Affects Version/s: Lustre 2.9.0
• Type: Bug
• Priority: Minor
• Reporter: John Hammond
• Assignee: Oleg Drokin
• Resolution: Unresolved
• Severity: 3

Description

t:lustre-release# export STRIPE_BYTES=$((1048576 * 4))
t:lustre-release# llmount.sh 
...
t:lustre-release# sleep 60 
t:lustre-release# lctl get_param lov.*.stripesize
lov.lustre-MDT0000-mdtlov.stripesize=4194304
lov.lustre-clilov-ffff88001c570be0.stripesize=1048576
t:lustre-release# cd /mnt/lustre
t:lustre# echo XXX > f0
t:lustre# lfs getstripe f0
f0
lmm_stripe_count:   1
lmm_stripe_size:    4194304
lmm_pattern:        1
lmm_layout_gen:     0
lmm_stripe_offset:  0
	obdidx		 objid		 objid		 group
	     0	             2	          0x2	             0

t:lustre# lctl get_param lov.*.stripesize
lov.lustre-MDT0000-mdtlov.stripesize=4194304
lov.lustre-clilov-ffff88001c570be0.stripesize=1048576

When we mount the client, we see the following on the console which is misleading and annoying:

[85604.363909] LustreError: 9299:0:(obd_config.c:1392:class_process_proc_param()) lov.: error writing proc entry 'stripesize': rc = -30

Issue Links

• is related to: LU-7334 Refactor lprocfs helpers

Comments

[Andreas Dilger]

My bad. This is related to patch http://review.whamcloud.com/20465 "LU-7334 lov: Cleanup lov_stripe proc files" because I thought those /proc files were left over from the lod/lov split. I think that patch should be reverted.

[Giuseppe Di Natale (Inactive)]

Should I abandon patch http://review.whamcloud.com//20465?

[James A Simmons]

If you revive the patch the LustreError reported needs to be addressed. Was the regression more than a annoying error message?