Description of problem:
Openshift people are looking for availability of thin_ls on rhel. This is
useful to figure out COW layer usage of container. Update device-mapper-persistent-data package to latest for 7.3.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
Here is upstream discussion looking for thin_ls.
This is needed for both Kubernetes and OpenShift. thin_ls gives us the ability to determine the amount of storage that a container has used in its COW layer.
Currently thin_ls does not operate on live metadata
$ ./thin_ls -o DEV,EXCLUSIVE /dev/mapper/fedora-docker--pool_tmeta
syscall 'open' failed: Device or resource busy
Note: you cannot run this tool with these options on live metadata.
This a problem for the cadvisor use case. Strangely, it seems to work when docker is using the loop thin dm rather than a real device.
Still looking into this.
In order for thin_ls to work on a live thin device a snapshot must be taken
dmsetup message /dev/mapper/fedora-docker--pool 0 "reserve_metadata_snap"
Then thin_ls will work. After thin_ls is run, the snapshot needs to be released
dmsetup message /dev/mapper/fedora-docker--pool 0 "release_metadata_snap"
# rpm -q device-mapper-persistent-data
# rpm -ql device-mapper-persistent-data | grep thin_ls
# lvcreate -L 200M -T vgtest/mythinpool -V1G -n thin1
# lvcreate -T vgtest/mythinpool -V1G -n thin2
# dmsetup message /dev/mapper/vgtest-mythinpool-tpool 0 reserve_metadata_snap
# thin_ls --metadata-snap /dev/mapper/vgtest-mythinpool_tmeta
DEV MAPPED CREATE_TIME SNAP_TIME
1 0 0 0
2 0 0 0
# dmsetup message /dev/mapper/vgtest-mythinpool-tpool 0 release_metadata_snap
*** Bug 1355797 has been marked as a duplicate of this bug. ***
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.
For information on the advisory, and where to find the updated
files, follow the link below.
If the solution does not work for you, open a new bug report.