Details
-
Bug
-
Resolution: Fixed
-
Critical
-
6.5.0
-
6.5.0-3840
-
Triaged
-
-
No
-
KV-Engine MH Beta part 2
Description
Steps to reproduce:
1. Initialize a 2 node cluster with default bucket having replica=1
+----------------+----------+--------------+
|
| Nodes | Services | Status |
|
+----------------+----------+--------------+
|
| 172.23.105.220 | [u'kv'] | Cluster node |
|
| 172.23.105.221 | None | <--- IN --- |
|
+----------------+----------+--------------+
|
2. Loaded 10k items into the bucket
+---------+---------+----------+-----+-------+------------+-----------+-----------+
|
| Bucket | Type | Replicas | TTL | Items | RAM Quota | RAM Used | Disk Used |
|
+---------+---------+----------+-----+-------+------------+-----------+-----------+
|
| default | membase | 1 | 0 | 10000 | 4477419520 | 141117168 | 76989548 |
|
+---------+---------+----------+-----+-------+------------+-----------+-----------+
|
3. Rebalance in 2 nodes and load 10k more items in parallel. Lots of DurabilityImpossibleExceptions are seen which is expected as per. Rebalance completed properly.
+----------------+----------+--------------+
|
| Nodes | Services | Status |
|
+----------------+----------+--------------+
|
| 172.23.105.220 | [u'kv'] | Cluster node |
|
| 172.23.105.221 | [u'kv'] | Cluster node |
|
| 172.23.105.223 | None | <--- IN --- |
|
| 172.23.105.225 | None | <--- IN --- |
|
+----------------+----------+--------------+
|
4. Repeat step 3. Rebalance in 1 more node and load 10k more docs.
Again many DurabilityImpossibleExceptions are seen.
+----------------+----------+--------------+
|
| Nodes | Services | Status |
|
+----------------+----------+--------------+
|
| 172.23.105.225 | [u'kv'] | Cluster node |
|
| 172.23.105.220 | [u'kv'] | Cluster node |
|
| 172.23.105.223 | [u'kv'] | Cluster node |
|
| 172.23.105.221 | [u'kv'] | Cluster node |
|
| 172.23.105.226 | None | <--- IN --- |
|
| 172.23.105.227 | None | <--- IN --- |
|
+----------------+----------+--------------+
|
Rebalance failed.
Attachments
Issue Links
- is duplicated by
-
MB-35270 [Durability] Rebalance stuck during replica update
- Closed
-
MB-35244 [Chain Rebalance]: Rebalance stuck while rebalancing in node with doc ops running in parallel having durability=majority
- Closed
-
MB-35274 Rebalance In 2 nodes is struck which were hard failover and moved out from the cluster earlier
- Closed