To use this hotfix (4.2.HF3).
Customers (except Canonical Openstack) and having Openstack Ussuri OR Openstack Victoria need to have an already deployed and working TVO-4.2 GA.
Customers (except Canonical Openstack) and having Openstack Wallaby need to follow the T4O-4.2 GA deployment process and directly upgrade to 4.2.HF3 containers/packages. The high-level flow below:
Deplo T4O-4.2 GA appliance.
Upgrade to 4.2.HF3 packages on the appliance.
Kolla
Deploy Trilio components via 4.2.HF3 containers/packages on Openstack Wallaby.
Openstack Ansible
Deploy Trilio components Openstack Wallaby [This will deploy 4.2 GA packages]
Upgrade TrilioVault packages to 4.2.HF3 on Openstack Wallaby.
Configure the Trilio appliance.
Canonical users having Openstack Ussuri Or Openstack Victoria can either upgrade (on top of 4.2 GA) using Trilio upgrade documents OR do a new deployment using 4.2 Deployment documents.
Canonical users having Openstack Wallaby need to do a new deployment using 4.2 Deployment documents.
Current Hotfix release targets the following:
High-level Qualification (via Sanity & Functional suites' execution) of T4O with Ussuri, Victoria & Wallaby Openstack.
Verification of Jira issues targeted for 4.2. release.
As part of the new process, the delivery will be via packages; end users would need to do the rolling upgrade on top of 4.2 GA.
Note : Container images with tag 4.2.64-hotfix-3-rhosp16.1 are not available for download from RedHat registry due to technical issues. Hence, it is recommended to use the latest tag, i.e. 4.2.64-hotfix-4-rhosp16.1.
Ref link for 4.2.64-hotfix-4-rhosp16.1 : T4O 4.2 HF4 Release Notes__
Artifacts
Reference
1
Release Date
Aug 25, 2022
2
Debian URL
deb [trusted=yes] https://apt.fury.io/triliodata-4-2/ /
3
RPM URL
4
PIP URL
****
Tag Reference in Install/Upgrade Docs
Value
Comments
1
4.2 Hotfix triliovault-cfg-scripts branch name
hotfix-3-TVO/4.2
Label against the Trilio repositories from where required code to be pulled for upgrades.
2
4.2 Hotfix RHOSP13 Container tag
4.2.64-hotfix-3-rhosp13
RHOSP13 Container tag for 4.2.HF3
3
4.2 Hotfix RHOSP16.1 Container tag
4.2.64-hotfix-3-rhosp16.1
RHOSP16.1 Container tag for 4.2.HF3
4
4.2 Hotfix RHOSP16.2 Container tag
4.2.64-hotfix-3-rhosp16.2
RHOSP16.2 Container tag for 4.2.HF3
5
4.2 Hotfix Kolla Victoria Container tag
4.2.64-hotfix-3-victoria
Kolla Container tag against 4.2.HF3
6
4.2 Hotfix Kolla Wallaby Container tag
4.2.64-hotfix-3-wallaby
Kolla Container tag against 4.2.HF3
7
4.2 Hotfix TripleO Container tag
4.2.63-hotfix-3-tripleo
TripleO Train CentOS 7 Container tag for 4.2.HF3
Summary
Restore failing while creating security group
Tvault configuration failing with build 4.1.19
Configuration fails with pcs auth
privsep Unhandled error: ConnectionRefusedError
Tvault configuration failing with build 4.1.19
Datamover container restarting
While deploying trilio-wlm 4.2 directly on the machine is getting stuck at workloadmgr package installation.
trilio data mover pods stuck in reboot loop post stack update on RHOSP 13
Reassigning a workload from a deleted project fails
Reassign of Workload from Deleted Project Fails SFDC #2821
default_tvault_dashboard_tvo-tvm not available after yum update
workload policy shows incorrect start time
tvault-config service is in the crash loop on 2 out of 3 nodes T4O cluster
Trilio core functionality operations do not perform as expected when the master T4O node is powered off.
backup stuck in uploading phase
Backup failed at snapshot_network_topology
Package/Container Names
Package Kind
Package Version/Container Tags
1
contego
deb
4.2.64
2
contegoclient
rpm
4.2.64-4.2
3
contegoclient
deb
4.2.64
4
contegoclient
python
4.2.64
5
dmapi
rpm
4.2.64-4.2
6
dmapi
deb
4.2.64
7
puppet-triliovault
rpm
4.2.64-4.2
8
python3-contegoclient
deb
4.2.64
9
python3-contegoclient-el8
rpm
4.2.64-4.2
10
python3-dmapi
deb
4.2.64
11
python3-dmapi
rpm
4.2.64-4.2
12
python3-s3-fuse-plugin
deb
4.2.64
13
python3-s3fuse-plugin
rpm
4.2.64-4.2
14
python3-trilio-fusepy
rpm
3.0.1-1
15
python-s3fuse-plugin-cent7
rpm
4.2.64-4.2
16
s3fuse
python
4.2.64
17
s3-fuse-plugin
deb
4.2.64
18
trilio-fusepy
rpm
3.0.1-1
19
4.2-RHOSP13-CONTAINER
Containers
4.2.64-hotfix-3-rhosp13
20
4.2-RHOSP16.1-CONTAINER
Containers
4.2.64-hotfix-3-rhosp16.1
21
4.2-RHOSP16.2-CONTAINER
Containers
4.2.64-hotfix-3-rhosp16.2
22
4.2-KOLLA-CONTAINER Victoria
Containers
4.2.64-hotfix-3-victoria
23
4.2-KOLLA-CONTAINER Wallaby
Containers
4.2.64-hotfix-3-wallaby
24
4.2-TRIPLEO-CONTAINER
Containers
4.2.64-hotfix-3-tripleo
Package/Container Names
Package Kind
Package/Container Version/Tags
1
python3-tvault-contego
deb
4.2.64.7
2
tvault-contego
deb
4.2.64.7
3
python3-tvault-contego
rpm
4.2.64.7-4.2
4
tvault-contego
rpm
4.2.64.7-4.2
5
workloadmgr
deb
4.2.64.6
6
workloadmgr
python
4.2.64.6
7
tvault_configurator
python
4.2.64.6
8
tvault-horizon-plugin
deb
4.2.64.1
9
tvault-horizon-plugin
rpm
4.2.64.1-4.2
10
python3-tvault-horizon-plugin
deb
4.2.64.1
11
python3-tvault-horizon-plugin-el8
rpm
4.2.64.1-4.2
12
python3-workloadmgrclient
deb
4.2.64.1
13
python3-workloadmgrclient-el8
rpm
4.2.64.1-4.2
14
python-workloadmgrclient
deb
4.2.64.1
15
workloadmgrclient
python
4.2.64.1
16
workloadmgrclient
rpm
4.2.64.1-4.2
Summary
Workaround/Comments (if any)
1
encrypted volume backup fails with SSO user
Follow below steps if T4O is reconfigured with ‘creator’ role
login to any T4O node
source particular user rc file
fire below command to get the trust id
workloadmgr trust-list 4. In order to create encrypted workload user needs to delete the existing trust which is created using other than ‘creator’ role
workloadmgr trust-delete <TrustID>
5.create a new trust with ‘creator’ role
workloadmgr trust-create creator
6.now create encrypted workload
2
additional security rule is getting added in shared security group after restore
It will go as known issue in 4.2HF3 and will be targeted in 4.2HF4
3
4
[encrypted] Post restore of incremental snapshots centos instance is not getting booted
There is no workaround as such. But customer can only restore the already taken full snapshot
5
[Intermittent] In-place restore doesn't work for ext3 & ext4 file system in canonical bionic-queens
In-place restore doesn't work well for ext3 & ext4 file system in canonical bionic-queens.
After in-place restore instance has data from the latest snapshot for ext3 & ext4 file system, however In-place restore was done for previous full/incremental snapshot.
6
Performance difference between encrypted & unencrypted WL/snapshot
With encryption in place, user would see some performance degradation against all operations done by Trilio.
Stats below as per trials in Trilio Lab
Snapshot time for LVM Volume Booted CentOS VM. Disk size 200 GB; total data including OS : ~108GB
For unencrypted WL : 62 min
For encrypted WL : 82 min
Snapshot time for Windows Image booted VM. No additional data except OS. : ~12 GB
For unencrypted WL : 10 min
For encrypted WL : 18 min
7
get-importworkload-list and get-orphaned-workloads-list are showing the wrong list of WLs
Customer need to use --project option with importworkload-lists cli to get the list WLs that can be imported with particular openstack.
workloadmgr workload-get-importworkloads-list --project_id <project_id>
8
File-search not displaying files present in logical vol on volume group (LVM)
If we create lvm partition using fdisk utility, then the file search will work.
9
Retention not working post snapshot mount/unmount operation
# List the Workload id for which Retention failing due to ownership change issue. # Fire the cmd chown -R nova:nova <WL_DIR> # After firing above cmd, now one should able to see the snapshot ownership as nova:nova.
10
[Barbican]File search on encrypted workload returns empty data
By default, if root directory is not having read permissions for group, then file search will also fail as it runs from nova user.
11
Single corrupted snapshot impacts import of all other valid snapshots causing file search failure
As per the current import design flow, if any single WL is corrupted (in current case few DB files were missing), then other good workloads get impacted during import, but import operation doesn’t stop OR fails. Respective wlm-api logs should show the error.
To mitigate the impact, the identified corrupted WL should be manually removed from target backend followed by reinitialize and import.
12
Test email error message should be in readable and understandable format
NA
13
File search will not work on Canonical if wlm is running on container (lxc container in this case)
NA
14
Unable to create encrypted workload if T4O reconfigured with creator trustee role.
If T4O is initially configured with member as trustee role and then user reconfigures the same with creator as a trustee role, then this failure would occur. Workaround : Follow below steps if T4O is reconfigured with ‘creator’ role
login to any T4O node
source particular user rc file
fire command to get the trust id (workloadmgr trust-list)
In order to create encrypted workload user needs to delete the existing trust which is created using other than ‘creator’ role (workloadmgr trust-delete <TrustID>)
create a new trust with ‘creator’ role (workloadmgr trust-create creator)
now create encrypted workload