doc(toolchain): sync loader with master - #478
Open
bitflicker64 wants to merge 1 commit into
Open
Conversation
Correct the loader mapping file and CLI reference against toolchain master: input source defaults and missing keys (has_header, dir_filter, kerberos_config, where, batch_size), the failure and progress file layout, the GRAPH input source rules, and the executor scripts. Adds sections for the flink-cdc loader and the spark-loader argument split, repairs a broken table row on the English page, and bumps the Docker image tags in the quick start examples from 1.5.0 to 1.7.0 so both language pages match the rest of the site.
bitflicker64
force-pushed
the
doc-sync/toolchain-loader
branch
from
September 5, 2026 18:57
2b331bf to
c13bce0
Compare
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Purpose of the PR
Sync the hugegraph-loader docs (en + cn) with hugegraph-toolchain master (3b385c3d).
The two quickstart pages had drifted from the code: several mapping-file defaults were wrong, a number of input-source keys were never documented, the breakpoint and failure-file section described a layout the loader no longer uses, and the flink-cdc loader had no section at all. The English page also had a table row with a heading glued onto it, which broke the rest of the parameter table.
The performance pages (
content/{en,cn}/docs/performance/hugegraph-loader-performance.md) state no options or defaults that master contradicts, so they are untouched.formatwas marked required with no defaultdelimiterwas documented as defaulting to a comma for every formathas_headerwas undocumentedextra_date_formatswas undocumenteddate_formatskipped_linewas said to skip nothing by default(^#|^//).*|and how to disable itlist_formatstart and end symbols were documented as[and]elem_delimitermust differ fromdelimiter, addedignored_elemscompressionentrycore_site_path, and did not mark it requiredhdfs_site_path,dir_filterandkerberos_configdriverwas marked requiredwherewas undocumentedbatch_size, no required keysbootstrap_server,topic,groupandformatrequired,from_beginningmapped to the offset reset,headerneeded for CSV and TEXT,delimiterused by TEXT only,batch_sizemapped tomax.poll.recordsgraphspacewas said to default toDEFAULT, and the mode and fallback rules were missinggraphspacemarked required, added the fallback to the matching command-line options, the rule that all input sources must then be GRAPH, the switch toRESTORINGmode, and the shape of the selected and ignored entriesskipnode was undocumented--max-connand--max-conn-per-routewere described only as something to tune by hand--batch-insert-threads--short-idhad no value formatlabel:field:typeand the accepted types--sink-typeand the HBase options carried no context--sink-typeto spark-loader, described both sinks, and marked the HBase options as belonging to--sink-type falseload-progress ${date}load-progress_yyyyMMdd-HHmmssand added that a resume reads the newest file in the directory--reload-failureand to.parse-errorand.insert-errorfiles under${struct}/current--failure-modeand for one${id}.errorfile per input source under${struct}/failure-data, with the tip-line format, the sibling${id}.header, the deletion of empty files, and append versus rewrite--max-read-errorswas missing from the list of limits that trigger a failure filebin/hugegraph-loaderbin/hugegraph-loader.shand documented how the script picks the JVM and passesJVM_OPTSutf8-bom-to-utf8.shwas not mentionedmapping-convert.shSPARK_HOMEandAPP_NAME, the exact option list that reaches the loader, the--filehandling under--deploy-mode cluster, the supported input sources, and the two sink modesFLINK_HOME, the MySQL JDBC mapping requirement and the two cdc options--batch-failure-fallbackrow had the 3.4.2 heading appended to it and the heading then appeared twice, so the table and the section both rendered wrongdocker runline and the docker-compose sample still pinned 1.5.0 images, while the rest of the site moved to 1.7.0 in the last content synchugegraph/loader,hugegraph/hugegraphandhugegraph/hubbleto 1.7.0, so the two pages agree with each other and with theVERSION=1.7.0download shown just below themrelease tagfor the stable version still gave 1.5.0 as its example