Guest

Preview Tool

Cisco Bug: CSCvs33041 - Write access on /mediaX is critical after upgrade to 7.14

Last Modified

Jan 24, 2020

Products (1)

  • Cisco Video Surveillance Manager

Known Affected Releases

7.14(1)

Description (partial)

Symptom:
/usr/BWhttpd/bin/reposdbtestclient dumpstoragemonshmem

returns zero values for filesystem stats on some partitions whereas return  correct ones for others

Volumes:
  vol #1:
    name:			/media1
    volume space (KB):		42949522432
    allocatable space (KB):	40802046310
    unallocated space (KB):	40802046310
    oversubscribed (KB):	0
    recording rate (Mbps):	0.000000
    num of archives:		0
    current state:		REPOS CLIPREPOS
    current status:		AVAILABLE
  vol #2:
    name:			/media2
    volume space (KB):		0
    allocatable space (KB):	0
    unallocated space (KB):	0
    oversubscribed (KB):	0
    recording rate (Mbps):	0.000000
    num of archives:		0
    current state:		REPOS CLIPREPOS
    current status:		UNAVAILABLE

ims.log errors

2019-12-04 15:45:15.306 [ SystemMonitor(17227) HEALTH_MON=1 <Sensor.cxx:67> ] Exception Caught: could not get fs status [/media2]
2019-12-04 15:45:15.311 [ SystemMonitor(17227) HEALTH_MON=1 <shmosutils.cxx:717> ] readfsusage: unable to  statfs() /media2/systemstoragestatus

Conditions:
I managed to reproduce that bug and it likely occurs if XFS filesystem was created on RHEL6 by using the manual method:

mkfs.xfs /dev/sdX

instead of running a script.

/usr/BWhttpd/bin/setup_media_storage

The reason behind is different i-node size used by default 

In RHEL6 
mkfs.xfs /dev/sdX

Creates 256 byte inodes by default

xfs_info /dev/sdd1
meta-data=/dev/sdd1              isize=256    agcount=40, agsize 268435455 blks

In RHEL7

mkfs.xfs /dev/sdX

Creates 512 byte inodes by default

xfs_info /dev/sdc1
meta-data=/dev/sdc1              isize=512    agcount=40, agsize=268435455 blks

RHEL6 or RHEL7
/usr/BWhttpd/bin/setup_media_storage

Creates 2048 byte inodes 
xfs_info /dev/sdb1
meta-data=/dev/sdb1              isize=2048   agcount=127, agsize=84546592 blks

I notice problem occurs only with isize=256 and 40TB or larger partitions (not tested for smaller but may work)

STEPS TO REPRODUCE:
+ create test 40TB disk and connect to VSM 7.11/7.12 VM
+ create GPT partition full size of a disk
+ create XFS filesystem manually using default values

mkfs.xfs /dev/sdX1

+ delete disk from 7.11/712 VM and connect to 7.14 VM
+ add /etc/fstab  entry for /dev/sdX1  in 7.14 VM
+ mount partition on /mediaX in 7.14 VM
+ enable /mediaX in 7.14 media server settings
+ run export LD_LIBRARY_PATH=/usr/BWhttpd/lib; /usr/BWhttpd/bin/reposdbtestclient dumpstoragemonshmem

QUICKER WAY:
+ Connect 40TB directly to 7.14 VM
+ create GPT partition full size off a disk
+ create a filesystem with i-node size 256 bytes

mkfs.xfs -f -i size=256 -m crc=0 /dev/sdX1

+ add /etc/fstab  entry for /dev/sdX1  in 7.14 VM
+ mount partition on /mediaX in 7.14 VM
+ enable /mediaX in 7.14 media server settings
+ run export LD_LIBRARY_PATH=/usr/BWhttpd/lib; /usr/BWhttpd/bin/reposdbtestclient dumpstoragemonshmem
Bug details contain sensitive information and therefore require a Cisco.com account to be viewed.

Bug Details Include

  • Full Description (including symptoms, conditions and workarounds)
  • Status
  • Severity
  • Known Fixed Releases
  • Related Community Discussions
  • Number of Related Support Cases
Bug information is viewable for customers and partners who have a service contract. Registered users can view up to 200 bugs per month without a service contract.