opencas service failed to run because of casctl settle failure

[liyimeng]

Description

In my system, I have open-cas service like this:

# /lib/systemd/system/open-cas.service
#
# Copyright(c) 2019-2021 Intel Corporation
# SPDX-License-Identifier: BSD-3-Clause
#

[Unit]
Description=opencas initialization service
After=systemd-remount-fs.service
Before=local-fs-pre.target local-fs.target
Wants=local-fs-pre.target local-fs.target
DefaultDependencies=no
OnFailure=emergency.target
OnFailureJobMode=isolate

[Service]
Type=oneshot
RemainAfterExit=yes
ExecStart=/sbin/casctl settle --timeout 1780 --interval 5
TimeoutStartSec=30min

[Install]
RequiredBy=local-fs.target local-fs-pre.target

# /etc/systemd/system/open-cas.service.d/override.conf
[Service]
ExecStartPre=/bin/bash -c 'echo Starting open-cas' 

but in systemd log, I see error like this

add_core() got an unexpected keyword argument 'try_add'

I try run the command manually, and get the same error

~# casctl settle 
add_core() got an unexpected keyword argument 'try_add'

So, it seems casctl settle is broken in some way

Expected Behavior

open-cas service should startup properly.

Actual Behavior

It failed

Steps to Reproduce

1. Install opencas 25.03
2. Configure opencas.conf
3. Retart OS
4. Error popup in systemd log

Context

Possible Fix

Logs

Configuration files

Your Environment

• OpenCAS version (commit hash or tag): 25.03
• Operating System: ubuntu
• Kernel version: 6.8.0
• Cache device type (NAND/Optane/other): nvme ssd
• Core device type (HDD/SSD/other): hdd
• Cache configuration:
• • Cache mode: (default: wt) wb
• • Cache line size: (default: 4) 32k
• • Promotion policy: (default: always)
• • Cleaning policy: (default: alru)acp
• • Sequential cutoff policy: (default: full)
• Other (e.g. lsblk, casadm -P, casadm -L)

[liyimeng]

I have created my cache instance as image below.

When I restart my node, I have to use casctl init --force to reactivate it, no other way at all! If I try to use casctl init I get

What is the actually problem! isn't that meta I am supposed to use. What is even ironic, I I want to clean up the meta data, is says that there is no meta data

What is the proper way to use opencas, driving me nuts! :(

[jfckm]

When I restart my node, I have to use casctl init --force to reactivate it, no other way at all! If I try to use casctl init I get

What is the actually problem! isn't that meta I am supposed to use. What is even ironic, I I want to clean up the meta data, is says that there is no meta data

In case where you already have metadata and want to start the cache casctl start is the proper command. casctl init is for initializing completely fresh instances and it fails if it detects old metadata unless you run it with --force.

[liyimeng]

@jfckm Thanks a lot! Your text is always helpful as usual :)

However, I think it is a bug in opencas that cause my problem. I have a PR for this, please help me to review #1699

[liyimeng]

Fixed by #1699