-
Type:
Bug
-
Resolution: Done
-
Priority:
Major - P3
-
None
-
Affects Version/s: None
-
Component/s: Aggregation Framework
-
Labels:None
-
ALL
-
I try to setup a backup procedure according to Back Up a Sharded Cluster with File System Snapshots
I have 3-member Config Replica Set with 4 shards. Each shard is a PSA Replica Set.
While backup my application is running of course. I run db.fsyncLock() on one Config SECONDARY. While the config RS member is locked my aggregation pipeline fails:
db.getSiblingDB('lau01mipmed03').getCollection('sessions#0.096-1500').aggregate([ { '$match': { 'n': { '$ne': 'dummy' }, 't': { '$gte': ISODate('2021-04-06T13:00:00.000Z'), '$lt': ISODate('2021-04-06T19:00:00.000Z') } } }, { '$unset': '_id' }, { '$merge': { 'into': { 'db': 'lau01mipmed03', 'coll': 'sessions#0' } } } ], { 'allowDiskUse': true })
Error:
"Request 13963276 timed out, deadline was 2021-04-06T15:00:42.947+02:00, op was RemoteCommand 13963276 -- target:[d-mipmdb-cfg-01.xxx.xxxx.xxx:27019] db:config expDate:2021-04-06T15:00:42.947+02:00 cmd:{ find: \"collections\", filter: { _id: \"lau01mipmed03.sessions#0\" }, readConcern: { level: \"majority\", afterOpTime: { ts: Timestamp(1617714012, 1), t: 2 } }, limit: 1, maxTimeMS: 30000 }"
Host d-mipmdb-cfg-01.xxx.xxxx.xxx is the locked Config SECONDARY.
Looks like, the error is not related to aggregation pipeline itself, the same error appears also on other commands, e.g.
db.getSiblingDB('t-mipmed-as-01').getCollection('sessions.096-1500.stats').aggregate([ { '$project': { 'stats': 1 } }, { '$unwind': '$stats' }, { '$replaceRoot': { 'newRoot': '$stats' } }, { '$merge': { 'into': { 'db': 'mip', 'coll': 'session.stats.raw' } } } ], { allowDiskUse: true })