When removing objects from a collection (many, many) I get this
message:
Wed Apr 18 17:43:58 [conn22408] old journal file will be removed: /var/
lib/mongo/journal/j._19
Wed Apr 18 17:43:58 [initandlisten] connection accepted from
10.1.0.2:57213 #22432
Wed Apr 18 17:43:59 [initandlisten] connection accepted from
10.1.0.2:53820 #22433
Wed Apr 18 17:43:59 [conn22408] old journal file will be removed: /var/
lib/mongo/journal/j._20
Wed Apr 18 17:43:59 [snapshotthread] cpu: elapsed:4004 writelock: 39%
For the next few minutes, my mongostat looks like this:
insert query update delete getmore command flushes mapped vsize
res faults locked % idx miss % qr|qw ar|aw netIn netOut conn
set repl time
0 0 0 0 0 3 0 56.2g 113g
9.08g 0 0 0 71|102 70|104 330b 1k 191
KNS M 17:53:46
0 0 0 0 0 1 0 56.2g 113g
9.08g 0 0 0 71|102 70|104 62b 1k 191
KNS M 17:53:47
0 0 0 0 0 4 0 56.2g 113g
9.08g 0 0 0 71|102 70|104 393b 2k 192
KNS M 17:53:48
0 0 0 0 0 1 0 56.2g 113g
9.08g 0 0 0 71|102 70|104 62b 1k 192
KNS M 17:53:49
Approximately when mongo starts responding, the log looks like this:
Wed Apr 18 17:49:08 [conn22408] DR101 latency warning on journal file
open 309182ms
Wed Apr 18 17:49:08 [conn22114] SocketException handling request,
closing client connection: 9001 socket exception [2] server
[
10.1.0.2:53241]
Wed Apr 18 17:49:08 [conn22428] SocketException handling request,
closing client connection: 9001 socket exception [2] server
[
10.1.0.2:57184]
..
Wed Apr 18 17:49:09 [conn22408] ClientCursor::find(): cursor not found
in map 8371761198641609067 (ok after a drop)
killcursors: found 0 of 1
Wed Apr 18 17:49:09 [conn22408] end connection XX.XX.XX.XX:45381
Does this happen every time a journal file is removed? How do I keep
this from happening?