From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from firstgate.proxmox.com (firstgate.proxmox.com [212.224.123.68]) by lore.proxmox.com (Postfix) with ESMTPS id 4C39D1FF13B for ; Wed, 25 Feb 2026 09:57:53 +0100 (CET) Received: from firstgate.proxmox.com (localhost [127.0.0.1]) by firstgate.proxmox.com (Proxmox) with ESMTP id 2C7E530BE9; Wed, 25 Feb 2026 09:58:47 +0100 (CET) Message-ID: Date: Wed, 25 Feb 2026 09:58:38 +0100 MIME-Version: 1.0 User-Agent: Mozilla Thunderbird Subject: Re: [pve-devel] [RFC PATCH-SERIES ha-manager 0/2] Negative Node Affinity Rules To: Daniel Kral , Proxmox VE development discussion References: <20251219133643.295514-1-d.kral@proxmox.com> <7dc9c2c9-e355-4a40-a00f-abdd9c81b9e2@proxmox.com> Content-Language: en-US From: Fiona Ebner In-Reply-To: Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit X-Bm-Milter-Handled: 55990f41-d878-4baa-be0a-ee34c49e34d2 X-Bm-Transport-Timestamp: 1772009906140 X-SPAM-LEVEL: Spam detection results: 0 AWL -1.075 Adjusted score from AWL reputation of From: address BAYES_00 -1.9 Bayes spam probability is 0 to 1% DMARC_MISSING 0.1 Missing DMARC policy KAM_DMARC_STATUS 0.01 Test Rule for DKIM or SPF Failure with Strict Alignment RCVD_IN_VALIDITY_CERTIFIED_BLOCKED 1.113 ADMINISTRATOR NOTICE: The query to Validity was blocked. See https://knowledge.validity.com/hc/en-us/articles/20961730681243 for more information. RCVD_IN_VALIDITY_RPBL_BLOCKED 0.358 ADMINISTRATOR NOTICE: The query to Validity was blocked. See https://knowledge.validity.com/hc/en-us/articles/20961730681243 for more information. RCVD_IN_VALIDITY_SAFE_BLOCKED 0.659 ADMINISTRATOR NOTICE: The query to Validity was blocked. See https://knowledge.validity.com/hc/en-us/articles/20961730681243 for more information. SPF_HELO_NONE 0.001 SPF: HELO does not publish an SPF Record SPF_PASS -0.001 SPF: sender matches SPF record Message-ID-Hash: YEFGUSWF5VHP66W4FTZPJH4STXKEEIJK X-Message-ID-Hash: YEFGUSWF5VHP66W4FTZPJH4STXKEEIJK X-MailFrom: f.ebner@proxmox.com X-Mailman-Rule-Misses: dmarc-mitigation; no-senders; approved; loop; banned-address; emergency; member-moderation; nonmember-moderation; administrivia; implicit-dest; max-recipients; max-size; news-moderation; no-subject; digests; suspicious-header X-Mailman-Version: 3.3.10 Precedence: list List-Id: Proxmox VE development discussion List-Help: List-Owner: List-Post: List-Subscribe: List-Unsubscribe: Am 25.02.26 um 9:45 AM schrieb Daniel Kral: > On Tue Feb 24, 2026 at 1:22 PM CET, Fiona Ebner wrote: >> Am 19.12.25 um 2:36 PM schrieb Daniel Kral: >>> For larger HA clusters, specifying the nodes in simple* node affinity >>> rules as opt-out (negative) instead of opt-in (positive) can make the >>> rule set easier to follow and implement by users. >>> >>> * simple = without priority groups >>> >>> >>> There's no web interface integration yet, because I'm not entirely sure >>> yet how to integrate it with the concept of priority groups for positive >>> node affinity rules, which do not make sense in this context as the >>> specified nodes will be removed from the effective node set. >> >> Wouldn't it be enough to not use/show the priority column when the >> affinity is negative? >> >> If people need both, to exclude certain nodes and to prioritize certain >> others, they can use two rules: >> 1. a negative node affinity rule >> 2. a non-strict positive node affinity rule with priorities > > Yes, this would only need some changes in the checks for node affinity > rules to allow both types for a HA resource at once and a check that > they do not contradict one another. > > I'm only contemplating about whether this could bite us later on if we > allow those rule sets and should wait for the demand for this. Waiting for demand is fine by me :) >>> As the conversion is pretty straightforward, we could even allow users >>> to convert between positive and negative node affinity rules (e.g. when >>> switching the affinity type in the web interface?). >> >> Limited to those without priorities I suppose ;) > > Right :). > > I'll see what makes sense in the web interface then, but it would > probably be the easiest to restrict changing the type of the node > affinity rule after it was created... Other options would be to either > just reset the node selection or drop the node priorities and invert the > selection if users change from 'positive' to 'negative'. Here too.