Details
-
Bug
-
Resolution: Fixed
-
Blocker
-
5.0.0
-
5.0.0-3456
-
Untriaged
-
-
Yes
Description
Could be duplicate of MB-25434 - if so please mark. I'm opening separate issue after speaking with Arun because we couldn't tell wether the rebalance was actually stuck or slowly progressing. The test is the same and repro/details are here: https://github.com/couchbaselabs/sequoia/blob/master/tests/integration/README.md
After 4 hours one of the vbuckets hasn't updated looking at updater_loop.
[ns_server:debug,2017-08-07T11:51:44.550-07:00,ns_1@172.23.108.103:<0.28713.0>:ns_rebalance_observer:docs_left_updater_loop:318]Starting docs_left_updater_loop:"default" |
[ns_server:debug,2017-08-07T11:51:49.551-07:00,ns_1@172.23.108.103:<0.28713.0>:ns_rebalance_observer:docs_left_updater_loop:318]Starting docs_left_updater_loop:"default" |
|
...
|
|
[ns_server:debug,2017-08-07T13:55:54.550-07:00,ns_1@172.23.108.103:<0.28713.0>:ns_rebalance_observer:docs_left_updater_loop:318]Starting docs_left_updater_loop:"default" |
[ns_server:debug,2017-08-07T13:55:59.550-07:00,ns_1@172.23.108.103:<0.28713.0>:ns_rebalance_observer:docs_left_updater_loop:318]Starting docs_left_updater_loop:"default" |
Seeing some I/O errors in memcached (172.23.97.237)
2017-08-07T11:46:04.793935-07:00 WARNING (WAREHOUSE) CouchKVStore::openDB: error:no such file [No such file or directory], name:/data/WAREHOUSE/270.couch.1, option:2, fileRev:1 |
2017-08-07T11:46:04.794116-07:00 WARNING (WAREHOUSE) doDcpVbTakeoverStats: exception while getting num persisted deletes for vbucket:270 - treating as 0 deletes. Details: CouchKVStore::getNumPersistedDeletes:Failed to open database file for vBucket = 270 rev = 1 with error:no such file: Input/output error |
We are not sure if the behavior is caused by addition of xattrs via sync_gateway to test. Next step is to re-run without the sg load. Unless there're any other recommendations.
Collect from 2 hrs ago attached, and updated collect is pending.