Skip to content

[DOC-14595][AI] Document disk_io_optimized for reader and writer threads - #4158

Merged
TimLFletcher merged 1 commit into
release/8.0from
DOC-14595
Sep 29, 2026
Merged

[DOC-14595][AI] Document disk_io_optimized for reader and writer threads#4158
TimLFletcher merged 1 commit into
release/8.0from
DOC-14595

Conversation

@TimLFletcher

@TimLFletcher TimLFletcher commented Sep 4, 2026 •
edited by atlassian Bot
Loading

Copy link
Copy Markdown
Contributor

Fixes DOC-14595.

What changed

modules/rest-api/pages/rest-reader-writer-thread-config.adoc (Setting Storage Thread Allocations).

num_reader_threads and num_writer_threads were described only as taking a number of threads. Neither mentioned disk_io_optimized, so a reader working from the REST reference had no way to discover it. Added the accepted value to both parameter descriptions and to the curl syntax block, with a link to Data Settings for guidance on when to use it.

Where the facts came from

Nothing here is invented. Every claim traces to the ticket or to existing pages in this repository:

Claim Source
disk_io_optimized is accepted by both parameters Ticket, including a working curl example
It allocates threads equal to the number of CPU cores manage/manage-settings/general-settings.adoc:224
Recommended for Magma buckets learn/buckets-memory-and-storage/storage-engines.adoc:49

I deliberately did not add default as a documented value. The UI has a Default option and the sibling num_nonio_threads parameter documents a default value, so it is probably accepted here too, but neither the ticket nor this repository states it. Worth a follow-up if someone can confirm.

Verification

Anchor [#data-settings] confirmed at general-settings.adoc:196, by inspection, since Antora validates xref page targets but not #fragment anchors.

PASS[docs-server @ DOC-14595]: no new build diagnostics.

🤖 Generated with Claude Code

@TimLFletcher @claude
[DOC-14595][AI] Document disk_io_optimized for reader and writer threads
acc98e7 
The Setting Storage Thread Allocations reference described
num_reader_threads and num_writer_threads as taking a number of threads,
without mentioning that they also accept disk_io_optimized. That value
corresponds to the Disk i/o optimized option already documented under Data
Settings, which storage-engines.adoc recommends for Magma buckets.

Adds the accepted value to both parameter descriptions and to the curl
syntax block, and links to Data Settings for guidance on when to use it.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
@TimLFletcher
TimLFletcher marked this pull request as ready for review September 29, 2026 14:45
@TimLFletcher
TimLFletcher merged commit c05cf25 into release/8.0 Sep 29, 2026
@TimLFletcher
TimLFletcher deleted the DOC-14595 branch September 29, 2026 14:45
pramodkumarka-sys pushed a commit that referenced this pull request Oct 14, 2026
@TimLFletcher @claude @pramodkumarka-sys
[DOC-14595][AI] Document disk_io_optimized for reader and writer thre…
507c327 
…ads (#4158)

The Setting Storage Thread Allocations reference described
num_reader_threads and num_writer_threads as taking a number of threads,
without mentioning that they also accept disk_io_optimized. That value
corresponds to the Disk i/o optimized option already documented under Data
Settings, which storage-engines.adoc recommends for Magma buckets.

Adds the accepted value to both parameter descriptions and to the curl
syntax block, and links to Data Settings for guidance on when to use it.

Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
(cherry picked from commit c05cf25)
pramodkumarka-sys added a commit that referenced this pull request Oct 15, 2026
@pramodkumarka-sys @RayOffiah
@TimLFletcher @claude @osfameron
[DOC-14793] Forward-port 16 release/8.0 doc fixes to release/8.5 (#4220)
d666fdd 
* [DOC-14713]: Feedback on Data | Couchbase Docs (#4179)

(cherry picked from commit b4f9cdb)

* [DOC-14595][AI] Document disk_io_optimized for reader and writer threads (#4158)

The Setting Storage Thread Allocations reference described
num_reader_threads and num_writer_threads as taking a number of threads,
without mentioning that they also accept disk_io_optimized. That value
corresponds to the Disk i/o optimized option already documented under Data
Settings, which storage-engines.adoc recommends for Magma buckets.

Adds the accepted value to both parameter descriptions and to the curl
syntax block, and links to Data Settings for guidance on when to use it.

Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
(cherry picked from commit c05cf25)

* [DOC-14444][AI] Add memory and XDCR criteria for choosing 1024 vBuckets (#4160)

Which Storage Engine Should You Use? did not mention two criteria that
point at the 1024 vBucket configuration: having at least 1 GiB of memory
per node available for the bucket, and needing XDCR replication from a
cluster older than 8.0.

Both are already documented elsewhere on the site. This surfaces them in
the decision list where a reader making the choice will see them.

Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
(cherry picked from commit 2432015)

* [DOC-14406][AI] Correct service flags in the update-services CLI syntax (#4161)

The couchbase-cli rebalance syntax on Modifying Services (and Rebalance)
listed --n1ql-[add|remove] and --cbas-[add|remove]. Neither is accepted by
the tool. The options are --query- and --analytics-.

The ticket reports the --n1ql- case. --cbas- is the same error on the same
line, corrected here as well.

Service names elsewhere on the page are unchanged: the service is still
called n1ql, only the command-line options differ.

Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
(cherry picked from commit 9ffc4ab)

* [DOC-14680][AI] Point cloud deployment links at the current Operator install page (#4163)

couchbase-cloud-deployment.adoc linked to operator:ROOT:install-gke.adoc,
install-eks.adoc and install-aks.adoc. All three were removed from
docs-operator by the K8S-1268 restructure.

They were consolidated into install-kubernetes.adoc, whose Certified
Kubernetes Platforms section lists EKS, GKE and AKS. All three links now
point there. Link text is unchanged, since it still describes the
destination.

Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
(cherry picked from commit 34cbf64)

* [DOC-14084][AI] Correct what the Read-Only Security Admin role can view (#4170)

The role was documented as unable to list users and groups. It can view
them, from 8.0.1. Corrected the section intro and the Security row of the
role table, which said "Cannot view or change users or groups".

The same page already contradicted this at the Read-Only Admin role, which
states that in 8.0 the permission to list users and groups was split off
into ro_security_admin.

Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
(cherry picked from commit eafe74b)

* [DOC-14002][AI] Say the vBucket count is fixed at bucket creation (#4185)

Edit a Bucket listed "Number of vBuckets" among the settings you can
change, saying you can change it for Magma buckets. The count is chosen at
creation and cannot be changed on an existing bucket.

Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
(cherry picked from commit ccbc88c)

* [DOC-14682][AI] List index-service streaming ports under encrypted node-to-node (#4186)

Ports 9100, 9101, 9103, 9104 and 9105 appeared only in the Unencrypted row
of the node-to-node summary. They have no separate TLS port and carry
encrypted traffic on the same port number, so anyone configuring a firewall
from that row closes them and breaks index builds.

Adds them to the Encrypted row and a note explaining that some node-to-node
ports have no separate TLS port. 9102 is unchanged: it does swap to 19102,
which is already listed.

Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
(cherry picked from commit c4b4353)

* [DOC-14275][AI] Correct the stated minimum for the rotateSize parameter (#4187)

The documented range read "11048576 to 524288000 (1 MiB to 500 MiB)". The
same sentence glosses the lower bound as 1 MiB, which is 1048576, so the
figure carried an extra leading digit and overstated the minimum by roughly
ten times.

Does not address the ticket's second request, that a value of 0 disables
size-based rotation. A comment on the ticket says 0 is accepted on 7.6 but
not on 8.0.

Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
(cherry picked from commit 58351fc)

* [DOC-14515][AI] Document dropUploadedCertificates on regenerateCertificate (#4188)

Regenerate All Certificates said that regeneration substitutes defaults for
the certificates an administrator uploaded, but did not mention the
dropUploadedCertificates parameter that prevents the uploaded node
certificates being dropped.

Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
(cherry picked from commit 920560e)

* [DOC-14514][AI] State that a cluster cannot span cloud providers (#4189)

Supported Platforms said clusters cannot mix operating systems, but said
nothing about cloud providers. Support had no documentation to point a
customer at when they ran a cluster across two providers.

Adds the statement to Supported Virtualization and Container Platforms.

Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
(cherry picked from commit dbd9a57)

* [DOC-14191][AI] Recommend storing data off the installation disk (#4190)

Sizing Guidelines listed disk recommendations but did not mention that
Couchbase Server writes data to the installation disk by default and that a
separate disk is preferable.

Added as a recommendation rather than a requirement, which is how the
ticket frames it.

Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
(cherry picked from commit 465848b)

* [DOC-14626][AI] Add the special characters introduced by MB-64667 (#4191)

Set Password Policy listed the pre-MB-64667 set of acceptable special
characters. Adds the eight symbols that MB-64667 identified as missing:
* = & . ; < > |

The existing entries and their order are unchanged.

Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
(cherry picked from commit eed23ff)

* [DOC-14652][AI] Document the client-certificate caveat for 7.6 to 8.0 upgrades (#4192)

Full mTLS for the Index Service arrived in 8.0 and is not supported on
mixed-version clusters. A rolling upgrade of a cluster with node-to-node
encryption and client-certificate authentication enabled fails the rebalance
that brings in the first upgraded Index node. Neither the upgrade guide nor
the client-certificate page mentioned it.

Adds a caveat and the disable/re-enable procedure to Before You Upgrade, and
a cross-reference from Client Certificates and Server Upgrade.

Does not add the release-note caveat the ticket also asks for.

Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
(cherry picked from commit 652532d)

* [DOC-14171][AI] State that mixed mode is unsupported outside an upgrade (#4193)

Upgrade referred to mixed mode in passing when describing finalization and
downgrade, but never said whether running a cluster that way is supported.
Adds the statement to Understanding Upgrade.

Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
(cherry picked from commit fa120b2)

* [DOC-14518][AI] Document internal client certificates (#4195)

Certificates covered root, node and client certificates but not the
internal client certificates a node presents when connecting to another
node. A customer deleted the default self-signed CA and invalidated theirs
with nothing in the docs explaining why.

Adds a section covering what they are, their requirements, and how to
upload custom ones.

Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
(cherry picked from commit 1e14e01)

* DOC-14793: Remove duplicate client-certificate section after cherry-pick of 652532d

* Doc 14675 fix Antora build errors (with Claude) (#4162)

* DOC-14675: Fix broken xrefs, includes, and an image path in server 8.0

Resolves a batch of Antora build errors surfaced for the server 8.0
component-version: typo'd link targets, `:` vs `/` module-separator
mistakes, `.html` used where `.adoc` was needed, bare same-directory
xrefs missing their subdirectory, a couple of browser "copy link to
highlighted text" fragments pasted in as xref targets, and one image
referenced from the wrong module.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

* DOC-14675: Fix four more broken xrefs missed in the first pass

- rest-bucket-create.adoc: two xrefs to xdcr-enable-crossclusterversioning.adoc
  were missing the `learn:` module prefix
- rotate-encryption-at-rest-key.adoc / drop-encryption-deks.adoc: xrefs used
  `manage:security/...` instead of `manage:manage-security/...`
- recover-data-with-xdcr.adoc: `add-a-node-and-rebalance.adoc` doesn't exist;
  real page is `add-node-and-rebalance.adoc`
- fts-system-indexes.adoc: xref used a filesystem-style path
  (`n1ql/pages/n1ql-language-reference/searchfun.adoc`) instead of proper
  xref syntax (`n1ql:n1ql-language-reference/searchfun.adoc`)

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

* DOC-14675: Fix duplicate doctitles and single-bracket section styles

- intro.adoc / starter-kits.adoc: each had a stray duplicate doctitle
  (level-0 heading) left over below the page attributes; removed the
  second one.
- rbac.adoc / authentication-domains.adoc: `[backup-restore]` and
  `[saml-authentication]` used single brackets, which Asciidoctor
  treats as a section *style* rather than an anchor - this makes the
  section a "special section" that can't contain nested subsections.
  Changed to double brackets (`[[id]]`) so they're proper anchors.
  authentication-domains.adoc's anchor is actively linked from
  authentication-overview.adoc and configure-saml.adoc, so it had to
  be preserved rather than deleted.

Verified locally with `asciidoctor` that both error types are gone
and no new warnings were introduced.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

* Delete orphaned pages with Asciidoc errors

* DOC-14675: Fix broken Sync Gateway cross-component xrefs (Batch 6)

Three xrefs into the sync-gateway component were broken:
- `sync-gateway:refer/config-properties.adoc` -- that path was renamed;
  `configuration-properties-legacy.adoc` carries a page-alias for it,
  confirming this is the real successor page, so linking directly to
  it rather than relying on alias resolution.
- Two uses of `sync-gateway::get-started-prepare.adoc` (double-colon,
  targeting the ROOT module) -- the real page lives in the
  `start-here` module, so the module name was missing.
- One of those also targeted `#step-1create-a-bucket`, an anchor that
  never existed on the target page. Added a real anchor
  (`#lbl-create-bucket`) to docs-sync-gateway itself, matching that
  page's existing anchor-naming convention, and pointed this xref at
  it instead of guessing.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

* DOC-14675: Fix stale swap-rebalance xref in recovery.adoc

Target page `install:upgrade-online.adoc` was deleted years ago.
Sibling pages (upgrade-cluster-online-full-capacity.adoc,
upgrade-cluster-online-reduced-capacity.adoc) already link to the
correct successor and anchor: upgrade-procedure-selection.adoc#swap-rebalance.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Sonnet 5 <noreply@anthropic.com>
(cherry picked from commit b8d6bc4)

* Review fixes for broken links

* DOC-14758: Remove broken cbdocloader nav entry (forward-port of 07afd95)

---------

Co-authored-by: Ray Offiah <77050471+RayOffiah@users.noreply.github.com>
Co-authored-by: Tim Fletcher <tim.fletcher@couchbase.com>
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
Co-authored-by: Hakim <hakim.cassimally@couchbase.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@supritha-kumar supritha-kumar supritha-kumar approved these changes

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

2 participants

@TimLFletcher @supritha-kumar