Details
-
Bug
-
Status: Closed
-
Major
-
Resolution: Fixed
-
2.6.0
-
Security Level: Public
-
CBL2.6.0-126 (net-msft Windows 10)
SG 2.5.1-11 (Centos 7)
CBS 6.0.1-2037 (Centos 7)
-
Medium
Description
Replication never finishes for test_replication_pull_filtering and test_replication_filter_access_revoke_document when filtering is done with 1000 docs. For 10 docs and 100 docs, test passes.
The issue is completed doesn't reaches to the exact count of expected total and after checking for sometime (150 iteration) we closes the replicator.
PS: This is with delta-sync disabled
pytest --timeout 3600 --liteserv-version=2.6.0-126 --liteserv-host=10.100.161.140 --liteserv-port=8080 --sg-ssl --enable-file-logging --no-conflicts --sync-gateway-version=2.5.1-11 --mode=cc --server-version=6.0.1-2037 --liteserv-platform=net-msft --create-db-per-test=cbl-test testsuites/CBLTester/CBL_Functional_tests/TestSetup_FunctionalTests -k test_replication_pull_filtering[1000]
As one can see because the replication never completes, we got key_error (additional field added in the doc) for the update on doc.
Seems like the status of the replicator is incorrect(i have manually added the progress of the replicator in the lite core, not part of default log).
CouchbaseLite Replicator Info: {Repl#8} pushStatus=stopped, pullStatus=busy, progress=42792/56992
If we write a unit test with a timeout ~100seconds, and replicate ~400 docs in which 200 should be rejected. then this can be reproduced. Find the attached unit-test.txt
and logs.txt
If i check the db elements count, i can see there are 201 documents. and progress as well.
cc. Jim Borden