Details
-
Bug
-
Resolution: Fixed
-
Test Blocker
-
3.0
-
Security Level: Public
-
None
-
Centos 6.4
Build 3.0.0-858-rel
-
Untriaged
-
Unknown
Description
1. Cluster : http://10.6.2.165:8091/
- 2 Buckets, 1 Ddoc,2 Views per bucket
- 120M, and 105M items on resp buckets
- 75% and 65% DGM on resp buckets
2. Load data and wait for indexing to complete.
3. Indexing for bucket with 120M items is stuck
- Indexing for the bucket with 120M items is stuck
- Seeing a large number of server timeouts on the nodes
"[couchdb:error,2014-06-23T19:08:54.321,ns_1@10.6.2.167:<0.2142.0>:couch_log:error:42]upr client (<0.2142.0>): Obtaining mutation from server timed out after 60.0 seconds [RequestId 9519 PartId 611]. Waiting..."
[root@centos-64-x64 logs]# grep -r "error" couchdb.log | wc
45 765 9450
- Cbstats shows no items are ready
./cbstats localhost:11210 upr | grep ready | grep true
eq_uprq:replication:ns_1@10.6.2.172->ns_1@10.6.2.167:default:stream_995_items_ready: false
eq_uprq:replication:ns_1@10.6.2.172->ns_1@10.6.2.167:default:stream_996_items_ready: false
eq_uprq:replication:ns_1@10.6.2.172->ns_1@10.6.2.167:default:stream_997_items_ready: false
eq_uprq:replication:ns_1@10.6.2.172->ns_1@10.6.2.167:default:stream_998_items_ready: false
Unacked bytes =0
[root@centos-64-x64 bin]# ./cbstats localhost:11210 upr | grep unack
eq_uprq:mapreduce_view: default _design/ddoc1 (prod/main):unacked_bytes: 0
eq_uprq:replication:ns_1@10.6.2.172->ns_1@10.6.2.163:default:unacked_bytes: 0
eq_uprq:replication:ns_1@10.6.2.172->ns_1@10.6.2.164:default:unacked_bytes: 0
eq_uprq:replication:ns_1@10.6.2.172->ns_1@10.6.2.165:default:unacked_bytes: 0
eq_uprq:replication:ns_1@10.6.2.172->ns_1@10.6.2.166:default:unacked_bytes: 0
eq_uprq:replication:ns_1@10.6.2.172->ns_1@10.6.2.167:default:unacked_bytes: 0
eq_uprq:replication:ns_1@10.6.2.172->ns_1@10.6.2.171:default:unacked_bytes: 0
Frequency of occurrence - Seen this on 2/2 runs so far with build 3.0.0-858-rel