Details
-
Bug
-
Resolution: Won't Fix
-
Major
-
2.2.0, 2.5.0, 3.0
-
Security Level: Public
-
Platform = Physical
OS = CentOS 6.5
CPU = Intel Xeon E5-2630
Memory = 64 GB
Disk = 2 x SSD
-
Centos 64-bit
-
Description
The test case is pretty simple:
1. Disable auto-compaction.
2. Create initial dataset, wait for persistence and TAP replication.
3. Compact bucket.
4. Define design documents.
5. Trigger index build, report total indexing time.
There is an issue with the last step. Usually it takes 25 minutes to finish initial indexing. However occasionally it takes longer, about 75 minutes with quite misleading progress reporting.
From logs I can see that 3 nodes finished indexing after 25 minutes:
[couchdb:info,2014-01-05T17:16:59.926,ns_1@172.23.96.15:<0.30369.0>:couch_log:info:39]Set view `bucket-1`, main group `_design/B`, updater finished
[couchdb:info,2014-01-05T17:18:01.914,ns_1@172.23.96.15:<0.30380.0>:couch_log:info:39]Set view `bucket-1`, main group `_design/D`, updater finished
[couchdb:info,2014-01-05T17:18:25.264,ns_1@172.23.96.15:<0.30348.0>:couch_log:info:39]Set view `bucket-1`, main group `_design/A`, updater finished
[couchdb:info,2014-01-05T17:18:33.827,ns_1@172.23.96.15:<0.30358.0>:couch_log:info:39]Set view `bucket-1`, main group `_design/C`, updater finished
[couchdb:info,2014-01-05T17:18:27.555,ns_1@172.23.96.17:<0.26484.0>:couch_log:info:39]Set view `bucket-1`, main group `_design/B`, updater finished
[couchdb:info,2014-01-05T17:19:23.092,ns_1@172.23.96.17:<0.26494.0>:couch_log:info:39]Set view `bucket-1`, main group `_design/D`, updater finished
[couchdb:info,2014-01-05T17:20:08.889,ns_1@172.23.96.17:<0.26474.0>:couch_log:info:39]Set view `bucket-1`, main group `_design/C`, updater finished
[couchdb:info,2014-01-05T17:20:37.673,ns_1@172.23.96.17:<0.26464.0>:couch_log:info:39]Set view `bucket-1`, main group `_design/A`, updater finished
[couchdb:info,2014-01-05T17:18:30.404,ns_1@172.23.96.18:<0.26487.0>:couch_log:info:39]Set view `bucket-1`, main group `_design/B`, updater finished
[couchdb:info,2014-01-05T17:20:28.915,ns_1@172.23.96.18:<0.26497.0>:couch_log:info:39]Set view `bucket-1`, main group `_design/D`, updater finished
[couchdb:info,2014-01-05T17:21:02.386,ns_1@172.23.96.18:<0.26477.0>:couch_log:info:39]Set view `bucket-1`, main group `_design/C`, updater finished
[couchdb:info,2014-01-05T17:21:32.953,ns_1@172.23.96.18:<0.26467.0>:couch_log:info:39]Set view `bucket-1`, main group `_design/A`, updater finished
But one of them finished indexing only after 75 minutes:
[couchdb:info,2014-01-05T18:06:44.109,ns_1@172.23.96.16:<0.26488.0>:couch_log:info:39]Set view `bucket-1`, main group `_design/C`, updater finished
[couchdb:info,2014-01-05T18:06:48.814,ns_1@172.23.96.16:<0.26508.0>:couch_log:info:39]Set view `bucket-1`, main group `_design/D`, updater finished
[couchdb:info,2014-01-05T18:06:53.955,ns_1@172.23.96.16:<0.26474.0>:couch_log:info:39]Set view `bucket-1`, main group `_design/A`, updater finished
[couchdb:info,2014-01-05T18:06:55.265,ns_1@172.23.96.16:<0.26498.0>:couch_log:info:39]Set view `bucket-1`, main group `_design/B`, updater finished
It apparently affects total indexing time and also makes aggregated progress a little bit confusing.
