df shows negative disk usage for a vxfs CFS mount point

book

Article ID: 100054437

calendar_today

Updated On:

Description

Error Message

None

Cause

Checking the superblock of the file system shows a very large number of free blocks. Typically more than are available in the file system. i.e.

echo "8192B.p S" | /opt/VRTS/bin/fsdb -t vxfs /dev/vx/rdsk/data02dg/vol01


super-block at 00000001.0000
magic a501fcf5  version 12 variant 1
uuid 6c6f405d-3007-0900-1678-0000ff9cb838
volguid 51f30fea-b2e6-11e9-b0cb-1a2b7b563d7c
ctime 1564503916 591664  (Tue Jul 30 12:25:16 2019 EDT)
log_version 13 logstart 0  logend 0
bsize  8192 size  264147712 dsize  264147712  ninode 0  nau 0
defiextsize 0  oilbsize 0  immedlen 96  ndaddr 10
aufirst 0  emap 0  imap 0  iextop 0  istart 0
bstart 0  femap 0  fimap 0  fiextop 0  fistart 0  fbstart 0
nindir 2048  aulen 32768  auimlen 0  auemlen 1
auilen 0  aupad 0  aublocks 32768  maxtier 15
inopb 32  inopau 0  ndiripau 0  iaddrlen 1   bshift 13
inoshift 5  bmask ffffe000  boffmask 1fff  checksum 1024bb3e5
free 140735597452917  ifree 0
efree  3019 2679 2695 2866 2801 2564 3448 2655 2099 1669 1212 896 595 423 257 4294908973 75 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 
flags 4000 mod 0 clean 69 pflags 1   

time 1624419126 924240  (Tue Jun 22 23:32:06 2021 EDT)
oltext[0] 11  oltext[1] 8220  oltsize 1
iauimlen 1  iausize 4  dinosize 256
checksum2 0  checksum3 0
fsetquotaction 0
fsetquotahardlimit 0 fsetquotasoftlimit 0
log_gen 13
fs_metadevid 0 fs_metablkno 0 fs_metatype 0
fs_bsoffset 0 fs_bsdevid 0 fs_bssize 0
fs_wb_loggen 0

In this output the number of free blocks far exceed the size (in blocks) of the file system. Another point of note is the pflags value of 1 shows the file system is clean and there is no corruption. 

Resolution

1. Unmount the file system on all nodes.

2. Even though the file system is marked claen, run fsck on the effected file system from one of the nodes. i.e.

fsck -t vxfs -y dev/vx/dsk/data02dg/vol01

3. Remount the file system.

4. To prevent this issue from re-occuring, contact Veritas Support and request the appropriate hotfix for your Infoscale Release referencing etrack 4090127. As of writing this fix is not available in any public patch. 

Issue/Introduction

df shows negative disk usage for a vxfs CFS mount point Example: $ df -h
/dev/vx/dsk/data02dg/vol01 2.0T -1.0E 1.0E - /var/opt/Storage02
$ df -hk
/dev/vx/dsk/data02dg/vol01 2113181696 -1125882606090352 1125884456657008 - /var/opt/Storage02