Add 'crm sbd' sub-level (jsc#PED-8256) #1491
Open
+2,773
−1,357
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
liangxin1300
force-pushed
the
20240614_crm_sbd_sublevel
branch
2 times, most recently
from
338ed50
to
2f10c6e
Compare
Disk-based SBD scenarios1. Show usage when syntax errorMore syntax errror cases 2. Completion3. Display SBD related configuration (UC4 in PED-8256)4. Change the on-disk meta data of the existing sbd disks (UC2.1 in PED-8256)5. Add a sbd disk with the existing sbd configuration (UC2.2 in PED-8256)6. Remove a sbd disk (UC2.3 in PED-8256)7. Purge sbd from cluster8. Replace the storage for a sbd disk (UC2.4 in PED-8256)9. display status (focusing on the runtime information only) (UC5 in PED-8256)10. overwrite case
Disk-less SBD scenarios1. Show usage when syntax error (diskless)2. completion (diskless)3. Display SBD related configuration (UC4 in PED-8256, diskless)4. Manipulate the basic diskless sbd configuration (UC3.1 in PED-8256)5. Remove diskless sbd from cluster |
liangxin1300
force-pushed
the
20240614_crm_sbd_sublevel
branch
9 times, most recently
from
a19a863
to
cc0d52a
Compare
liangxin1300
force-pushed
the
20240614_crm_sbd_sublevel
branch
9 times, most recently
from
1456931
to
e8f53af
Compare
liangxin1300
force-pushed
the
20240614_crm_sbd_sublevel
branch
2 times, most recently
from
bc2a1fa
to
229de46
Compare
liangxin1300
force-pushed
the
20240614_crm_sbd_sublevel
branch
6 times, most recently
from
77c1c4f
to
5d17668
Compare
liangxin1300
force-pushed
the
20240614_crm_sbd_sublevel
branch
from
5d17668
to
ce84f84
Compare
…rties under diskless sbd
After adding sbd device interface to manage devices, related functionalities inside sbd configure interface should be adjusted
and make sure the metadata is consistent between devices.
Add a log message to indicate the start of pacemaker.service. This helps users understand that the system is not hanging but is actually starting pacemaker, especially when SBD_DELAY_START is set and it takes longer to start pacemaker.
to avoid duplicate info message.
to redirect stderr to stdout.
And the `sbd purge` command will also move /etc/sysconfig/sbd to /etc/sysconfig/sbd.bak on all nodes.
liangxin1300
force-pushed
the
20240614_crm_sbd_sublevel
branch
2 times, most recently
from
774ea69
to
79535f6
Compare
zzhou1
reviewed
Nov 28, 2024
zzhou1
reviewed
Nov 28, 2024
liangxin1300
force-pushed
the
20240614_crm_sbd_sublevel
branch
from
79535f6
to
2cbfc71
Compare
|
Add output of sbd process in |
- Return immediately if no changes are made - Adjust watchdog timeout and msgwait values properly
liangxin1300
force-pushed
the
20240614_crm_sbd_sublevel
branch
2 times, most recently
from
5a33305
to
b9e9853
Compare
zzhou1
reviewed
Nov 29, 2024
doc/crm.8.adoc
Outdated
| ............... | ||
| # For disk-based SBD | ||
| crm sbd configure show [disk_metadata|sysconfig|property] | ||
| crm sbd configure [device=<dev>]... [watchdog-device=<dev>] [watchdog-timeout=<integer>] [allocate-timeout=<integer>] [loop-timeout=<integer>] [msgwait-timeout=<integer>] |
There was a problem hiding this comment.
And I am confused by
crm sbd configure watchdog-timeout=.... There are 3 similar items:Timeout (watchdog) :,SBD_WATCHDOG_TIMEOUT=andstonith-watchdog-timeout=. Which ones are expected be modified by this command?
Two major scenarios:
- disk-based
Timeout (watchdog) :in the disk metadata is used.SBD_WATCHDOG_TIMEOUT=is useless - diskless
SBD_WATCHDOG_TIMEOUT=andstonith-watchdog-timeout=are meant to be used by diskless-sbd only
doc/crm.8.adoc
Outdated
| ............... | ||
| # For disk-based SBD | ||
| crm sbd configure show [disk_metadata|sysconfig|property] | ||
| crm sbd configure [device=<dev>]... [watchdog-device=<dev>] [watchdog-timeout=<integer>] [allocate-timeout=<integer>] [loop-timeout=<integer>] [msgwait-timeout=<integer>] |
There was a problem hiding this comment.
If
Timeout (watchdog) :andSBD_WATCHDOG_TIMEOUT=controls the same thing, and only one of them is effective, we should show only the effective one incrm sbd configure show, or indicate which one is effective in some way.
I'm kind of agree with you. Let's keep debating with Xin ;)
| # To keep the order of devices during removal | ||
| left_device_list = [dev for dev in self.device_list_from_config if dev not in devices_to_remove] | ||
| if len(left_device_list) == 0: | ||
| raise self.SyntaxError("Not allowed to remove all devices") |
There was a problem hiding this comment.
Suggested change
| raise self.SyntaxError("Not allowed to remove all devices") | |
| raise self.SyntaxError("Not allowed to remove all devices. Run `crm cluster init sbd -S` to bootstrap the diskless-sbd") |
Intentionally, not give "-F" directly here, to have user think this twice. We can debate this.
liangxin1300
force-pushed
the
20240614_crm_sbd_sublevel
branch
from
b9e9853
to
1dedfe3
Compare
liangxin1300
force-pushed
the
20240614_crm_sbd_sublevel
branch
from
1dedfe3
to
a9636b9
Compare
zzhou1
reviewed
Nov 29, 2024
| return False | ||
| if not sbd.SBDUtils.is_using_disk_based_sbd(): | ||
| logger.error("Only works for disk-based SBD") | ||
| logger.info("Please use 'crm cluster init -s <dev1> [-s <dev2> [-s <dev3>]]' to configure disk-based SBD first") |
There was a problem hiding this comment.
Probably better to suggest using the SBD stage
Suggested change
| logger.info("Please use 'crm cluster init -s <dev1> [-s <dev2> [-s <dev3>]]' to configure disk-based SBD first") | |
| logger.info("Please use 'crm cluster init sbd -s <dev1> [-s <dev2>]' to configure the disk-based SBD first") |
| for node in self.cluster_nodes: | ||
| out = self.cluster_shell.get_stdout_or_raise_error(scripts_in_shell, node) | ||
| if out: | ||
| print(f"# Status of sbd process on {node}:") |
There was a problem hiding this comment.
Suggested change
| print(f"# Status of sbd process on {node}:") | |
| print(f"# Status of the sbd disk watcher process on {node}:") |
And, this information should no be printed out for the diskless-sbd.
| ) | ||
| sbd_manager.init_and_deploy_sbd() | ||
|
|
||
| def _configure_diskless(self, parameter_dict: dict): |
There was a problem hiding this comment.
Same as the disk-based sbd, I expect to see TimeoutStartSec get updated along with crm sbd configure watchdog-timeout=50, for example.
Sounds like, the diskless bootstrap code doesn't do this too.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Motivation
The main configurations for sbd use cases are scattered among sysconfig,
on-disk meta data, CIB, and even could be related to other OS components
eg. coredump, SCSI, multipath.
It's desirable to reduce the management complexity among them and to
streamline the workflow for the main use case scenarios.
Changed include
Disk-based SBD scenarios
Disk-less SBD scenarios