ondrejhome / ansible.ha-cluster-lvm Goto Github PK

Issue:
default value below is broken
local_volume_list: "{% for i in ansible_lvm['vgs'] %}\"{{ i }}\", {% endfor %}"
If the playbook is run more than once, this gets populated with VGS that are visible and this can include shared VGs which is undesirable to appear in the volume_list

Possible solutions:

• specifying a filter to exclude the shared VGs ?
• dropping this default and requiring manual specification?

In RHEL8 environment, HALVMtype is set to systemid. And when I run role, it fails when creating lv.
It is an iSCSI-connected volume and is allocated as sdb.

Below is the failure message that occurs when performing ansible.
{
"ansible_loop_var": "item",
"_ansible_item_label": [
{
"pvs": "/dev/sdb",
"name": "HAvg01"
},
{
"path": "/hatest1",
"name": "lv_hatest01",
"size": "10g"
}
],
"err": " Volume "HAvg01/lv_hatest01" is not active locally (volume_list activation filter?).\n Aborting. Failed to wipe start of new LV.\n",
"rc": 5,
"item": [
{
"pvs": "/dev/sdb",
"name": "HAvg01"
},
{
"path": "/hatest1",
"name": "lv_hatest01",
"size": "10g"
}
],
"msg": "Creating logical volume 'lv_hatest01' failed",
"changed": false,
"invocation": {
"module_args": {
"pvs": null,
"force": false,
"vg": "HAvg01",
"lv": "lv_hatest01",
"resizefs": false,
"state": "present",
"thinpool": null,
"snapshot": null,
"active": true,
"shrink": true,
"opts": null,
"size": "10g"
}
},
"_ansible_no_log": false
}

To simplify the code the rgmanager-only version of this roles was created in https://github.com/OndrejHome/ansible.ha-cluster-lvm-rgmanager

Idea is to remove rgmanager code from this repository and keep only pacemaker related code.

TODO:

• remove all rgmanager related code
• re-test on EL6 pacemaker platform (at least CentOS 6.10 and RHEL 6.10) to confirm that this works for both tagging and clvm there
• implement repository detection (from ondrejhome.ha-cluster.pacemaker code)
• update documentation on examples and provide link to rgmanager role

lvol module issue.

TASK [ondrejhome.ha-cluster-lvm : create systemid VG] *********************************************************************************************************
changed: [192.168.22.58] => (item={'name': 'vg_systemid', 'pvs': '/dev/sdb', 'lvs': [{'name': 'lv_name1', 'size': '200'}, {'name': 'lv_name2', 'size': '100'}]})
 [WARNING]: The value 4 (type int) in a string field was converted to '4' (type string). If this does not look like what you expect, quote the entire value to
ensure it does not change.

Idea: Implement lvm2-lockd configuration for RHEL 8 as described in docs for GFS2 implementation - https://access.redhat.com/documentation/en-us/red_hat_enterprise_linux/8/html/configuring_and_managing_high_availability_clusters/assembly_configuring-gfs2-in-a-cluster-configuring-and-managing-high-availability-clusters#proc_configuring-gfs2-in-a-cluster.adoc-configuring-gfs2-cluster

potential issues:

• pcs-modules-2 and the cloned resource group support.

After creating a LVM-activate resource to manage a systemid VG, the resource starts without problems on the node where the VG was created. However, when it needs to fail over to another node, it fails with the following error during the start of the resource:

Failed Resource Actions:
  * my_lvm start on fastvm-rhel-9-1-65 returned 'error' (Volume group[vg_shared] doesn't exist, or not visible on this node!) at Sat Nov 19 16:45:24 2022 after 234ms