Description of problem: ----------------------- When creating thick LVs on top of VDO volume in certain situations, 'insufficient space' is observed. The situations are: 1. Create only one brick on a disk and enable dedupe & compression Eg: 'data' volume with brick size of 12TB on disk 'sdb' and enable dudupe & compression 2. Create one or more bricks and enable dedupe & compression, as well as create a separate brick on another disk Eg: 'data' volume with brick size of 12TB on disk 'sdb', enable dedupe & compression 'vmstore' volume with brick size of 4TB on the disk 'sdb', enable dedupe & compression 'extravol' volume with brick size of 1TB on the disk 'sdc' Version-Release number of selected component (if applicable): ------------------------------------------------------------- cockpit-ovirt-dashboard-0.12.5 How reproducible: ----------------- Always Steps to Reproduce: ------------------- Method 1: 1. From cockpit HC deployment wizard, create a volume, for example: 'data' volume as replica type 2. Select one full disk for the brick on each host, for example: sdc 3. Provide the size of the brick. Example: 12288GB ( 12TB ) 4. Enable dedupe & compression Method 2: 1. From cockpit HC deployment wizard, create 3 volumes. Example:'data' volume as replica type, 'vmstore' volume as replica type, 'extravol' as replica type 2. Create 2 bricks on the same disk. For example, brick of 'vmstore' & 'data' resides on 'sdc' 3. Enable dedupe and compression on that disk 4. Place the brick of another volume on a different disk and do not enable dedupe or compression. For example: brick of extravol is placed on 'sdd' Actual results: --------------- Insufficient space in volume group while creating the thick LV Expected results: ----------------- thick LV creation should be successful
Created attachment 1545920 [details] generated inventory file This is the inventory file generated for: 1. 4 volumes ( engine, vmstore, data, extravol ) 2. Bricks placed on sdb, sdc, sdc, sdd respectively
Tested with cockpit-ovirt-dashboard-0.12.6 1. When only one brick is created out of disk, the brick is created with size value as 100%FREE <snip> gluster_infra_vdo: - name: vdo_sdc device: /dev/sdc slabsize: 32G logicalsize: 8000G blockmapcachesize: 128M readcache: enabled readcachesize: 20M emulate512: 'on' writepolicy: auto gluster_infra_thick_lvs: - vgname: gluster_vg_sdb lvname: gluster_lv_engine size: 100G - vgname: gluster_vg_sdc lvname: gluster_lv_data size: 100%FREE </snip> 2. When many bricks ( with VDO ) are created on the same disk, then also the size of the last thick LV is set to 100%FREE <snip> gluster_infra_thick_lvs: - vgname: gluster_vg_sdb lvname: gluster_lv_engine size: 100G - vgname: gluster_vg_sdc lvname: gluster_lv_data size: 5000G - vgname: gluster_vg_sdc lvname: gluster_lv_vmstore size: 100%FREE </snip>
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. https://access.redhat.com/errata/RHEA-2019:1121