From: Daniel Kral <d.kral@proxmox.com>
To: pve-devel@lists.proxmox.com
Subject: [pve-devel] [PATCH ha-manager v3 21/21] rules: add documentation about current feasibility check implementations
Date: Mon,  3 Nov 2025 11:20:05 +0100	[thread overview]
Message-ID: <20251103102118.153666-22-d.kral@proxmox.com> (raw)
In-Reply-To: <20251103102118.153666-1-d.kral@proxmox.com>
These notes should clarify the goals and reasons for the feasibility
checks while giving direction how these can be improved in the future.
Signed-off-by: Daniel Kral <d.kral@proxmox.com>
---
changes since v2:
  - NEW!
 src/PVE/HA/Rules.pm | 19 ++++++++++++++++---
 1 file changed, 16 insertions(+), 3 deletions(-)
diff --git a/src/PVE/HA/Rules.pm b/src/PVE/HA/Rules.pm
index 9d7f12a1..c4a2ccea 100644
--- a/src/PVE/HA/Rules.pm
+++ b/src/PVE/HA/Rules.pm
@@ -32,6 +32,13 @@ the feasibility between rules of the same type and and between rules of
 different types, and prune the rule set in such a way, that it becomes feasible
 again, while minimizing the amount of rules that need to be pruned.
 
+The feasibility checks must verify that all rules can be applied error-free in
+any runtime state of the HA resources, cluster nodes, and the HA Manager in a
+reasonable amount of time. This might require more rules to be rejected than
+necessary as some checks can become computationally expensive for little benefit
+to the core feature, but this can certainly be reevaluated if there are user
+requests for these cases.
+
 More so, the rules given by the config file might not be in the best format to
 be used internally or does not contain the implicitly stated rules, which are
 induced by the relationship between different rules. Therefore, this package
@@ -598,9 +605,11 @@ a list of lists, each consisting of the rule type and rule id, where at
 least one resource in a resource affinity rule is in a node affinity rule,
 which has multiple priority groups defined.
 
-That is, the resource affinity rule cannot be statically checked to be feasible
-as the selection of the priority group is dependent on the currently online
-nodes.
+These cases are currently rejected, because the selection of the highest
+priority class is dependent on the currently online nodes and on the current
+HA resource states and node placements in the case of negative resource affinity
+rules. This makes the verification of these rules' feasibility in all possible
+states computationally hard.
 
 If there are none, the returned list is empty.
 
@@ -670,6 +679,10 @@ lists, each item consisting of the rule type and rule id, where at least one of
 the resources is used in a positive resource affinity rule and more than one
 node affinity rule.
 
+These cases are currently rejected, because there is no definitive method to
+merge multiple node affinity rules, e.g. with different strictness values and
+different priority classes, yet.
+
 If there are none, the returned list is empty.
 
 =cut
-- 
2.47.3
_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
next prev parent reply	other threads:[~2025-11-03 10:23 UTC|newest]
Thread overview: 24+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-11-03 10:19 [pve-devel] [PATCH ha-manager v3 00/21] HA rules fixes + performance improvements + cleanup Daniel Kral
2025-11-03 10:19 ` [pve-devel] [PATCH ha-manager v3 01/21] config: do not add ignored resources to dependent resources Daniel Kral
2025-11-03 10:19 ` [pve-devel] [PATCH ha-manager v3 02/21] manager: retranslate rules if nodes are added or removed Daniel Kral
2025-11-03 10:19 ` [pve-devel] [PATCH ha-manager v3 03/21] rules: factor out disjoint rules' resource set helper Daniel Kral
2025-11-03 10:19 ` [pve-devel] [PATCH ha-manager v3 04/21] rules: resource affinity: inter-consistency check with merged positive rules Daniel Kral
2025-11-03 10:19 ` [pve-devel] [PATCH ha-manager v3 05/21] rules: add merged positive resource affinity info in global checks Daniel Kral
2025-11-03 17:24   ` Michael Köppl
2025-11-03 10:19 ` [pve-devel] [PATCH ha-manager v3 06/21] rules: make rules sorting optional in foreach_rule helper Daniel Kral
2025-11-03 10:19 ` [pve-devel] [PATCH ha-manager v3 07/21] rename rule's canonicalize stage to transform stage Daniel Kral
2025-11-03 10:19 ` [pve-devel] [PATCH ha-manager v3 08/21] rules: make plugins register transformers instead of plugin_transform Daniel Kral
2025-11-03 10:19 ` [pve-devel] [PATCH ha-manager v3 09/21] rules: node affinity: decouple get_node_affinity helper from Usage class Daniel Kral
2025-11-03 10:19 ` [pve-devel] [PATCH ha-manager v3 10/21] compile ha rules to a more compact representation Daniel Kral
2025-11-03 10:19 ` [pve-devel] [PATCH ha-manager v3 11/21] test: rules: use to_json instead of Data::Dumper for config output Daniel Kral
2025-11-03 10:19 ` [pve-devel] [PATCH ha-manager v3 12/21] test: rules: add compiled config output to rules config test cases Daniel Kral
2025-11-03 10:19 ` [pve-devel] [PATCH ha-manager v3 13/21] rules: node affinity: define node priority outside hash access Daniel Kral
2025-11-03 10:19 ` [pve-devel] [PATCH ha-manager v3 14/21] move minimum version check helper to ha tools Daniel Kral
2025-11-03 10:19 ` [pve-devel] [PATCH ha-manager v3 15/21] manager: move group migration cooldown variable into helper Daniel Kral
2025-11-03 10:20 ` [pve-devel] [PATCH ha-manager v3 16/21] api: status: sync active service counting with lrm's helper Daniel Kral
2025-11-03 10:20 ` [pve-devel] [PATCH ha-manager v3 17/21] manager: group migration: " Daniel Kral
2025-11-03 10:20 ` [pve-devel] [PATCH ha-manager v3 18/21] factor out counting of active services into helper Daniel Kral
2025-11-03 10:20 ` [pve-devel] [PATCH ha-manager v3 19/21] tree-wide: remove misused function prototype declaractions Daniel Kral
2025-11-03 10:20 ` [pve-devel] [PATCH ha-manager v3 20/21] rules: fix documentation for inter-rules checker subroutines Daniel Kral
2025-11-03 10:20 ` Daniel Kral [this message]
2025-11-03 17:43 ` [pve-devel] [PATCH ha-manager v3 00/21] HA rules fixes + performance improvements + cleanup Michael Köppl
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox
  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):
  git send-email \
    --in-reply-to=20251103102118.153666-22-d.kral@proxmox.com \
    --to=d.kral@proxmox.com \
    --cc=pve-devel@lists.proxmox.com \
    /path/to/YOUR_REPLY
  https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
  Be sure your reply has a Subject: header at the top and a blank line
  before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.