Details
-
Bug
-
Status: Closed
-
Critical
-
Resolution: Fixed
-
6.6.0
-
Enterprise Edition 6.6.0 build 7883/6.6.0-7897
-
Triaged
-
Centos 64-bit
-
-
1
-
Yes
Description
– Create a replication between default bucket on source cluster to remove bucket on destination cluster. (Default bucket on source cluster has - compression - Active)
– Following error is seen in babysitter.log and memcached.log. CPU on both the nodes is at 90%. Destination nodes have 2 nodes in cluster.
Note - the same test on another cluster run smoothly without the error above.
2020-07-26T20:55:59.064767-07:00 ERROR 319: exception occurred in runloop during packet execution. Cookie info: [{"aiostat":"success","connection":"[ 172.23.108.103:44786 - 172.23.105.111:11210 (<ud>Administrator</ud>) ]","engine_storage":"0x0000000000000000","ewouldblock":false,"packet":{"bodylen":402,"cas":0,"datatype":["Snappy","Xattr"],"extlen":24,"key":"<ud>9DC5E3F5-22_1536349</ud>","keylen":19,"magic":"ClientRequest","opaque":251658240,"opcode":"DEL_WITH_META","vbucket":179},"refcount":0}] - closing connection ([ 172.23.108.103:44786 - 172.23.105.111:11210 (<ud>Administrator</ud>) ]): get_body_size: Failed to inflate data
|
2020-07-26T20:55:59.065379-07:00 WARNING (remote) Invalid streaming connection
|
From the log it looks like we're trying to inflate an item which claims to be compressed and contains xattrs, but we fail to inflate it and my guess is that the document isn't compressed with snappy.
Could this be introduced with the changes we did for transactions?