flimzy commented on a change in pull request #348: Improve clustered purge documentation URL: https://github.com/apache/couchdb-documentation/pull/348#discussion_r233371243
########## File path: src/cluster/purging.rst ########## @@ -129,45 +131,52 @@ consistency within the main database. Config Settings ==================================== -These settings ca be updated in the default.ini or local.ini: +These settings can be updated in the default.ini or local.ini: + ++-----------------------+--------------------------------------------+----------+ +| Field | Description | Default | ++=======================+============================================+==========+ +| max_document_id_number| Allowed maximum number of documents in one | 100 | +| | purge request | | ++-----------------------+--------------------------------------------+----------+ +| max_revisions_number | Allowed maximum number of accumulated | 1000 | +| | revisions in one purge request. | | ++-----------------------+--------------------------------------------+----------+ +| allowed_purge_seq_lag | Beside purged_infos_limit, allowed | 100 | +| | additional buffer to store purge requests | | ++-----------------------+--------------------------------------------+----------+ +| index_lag_warn_seconds| Allowed durations when index is not | | +| | updated for local purge checkpoint document| 86400 | ++-----------------------+--------------------------------------------+----------+ + +During a database compaction, we check all checkpoint purge docs. A client (an +index or internal replication job) is allowed to have the last reported +``purge_seq`` to be smaller than the current database shard's ``purge_seq`` by +the value of ``(purged_infos_limit + allowed_purge_seq_lag)``. If the client's +``purge_seq`` is even smaller, and the client has not checkpointed within +``index_lag_warn_seconds``, it prevents compaction of purge trees and we have to +issue the following log warning for this client: + +.. code-block:: text + + Purge checkpoint '_local/purge-mrview-9152d15c12011288629bcffba7693fd4’ + not updated in 86400 seconds in + <<"shards/00000000-1fffffff/testdb12.1491979089">> + +If this type of log warnings occurs, check the client to see why the processing +of purge requests is stalled in it. + +There is a mapping relationship between design document of indexes and local +checkpoint docs. If a design document of indexes is updated or deleted, the +corresponding local checkpoint document should be also automatically deleted. +But in an unexpected case, when a design doc was updated/deleted, but its +checkpoint document still exist in a database, the following warning will be Review comment: ```suggestion checkpoint document still exists in a database, the following warning will be ``` ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
