1

When using MMAPv1, the storage location of the database(s) is easily identified. For example,the "test" database can easily be identified here:

"/data/db/test.ns"

In this way, you are assured that this is where the database lives. This can be very helpful to know; like helping to ensure you are backing up the correct databases as one example. Confirming that restored data is where it is expected to be is another.

I'd like to know if I can also identify the database names when the Storage Engine is Wired Tiger. I have seen the format of 

"collection-000-*.wt"

but, this does not allow me to see the actual names of the collections or the databases they live in.

I appreciate any help on this. Thanks, Rick Farina

rickfarina
  • 116
  • 1
  • 3
  • https://docs.mongodb.org/v3.0/reference/configuration-options/#storage.directoryPerDB might help you here. However, I fail to see how knowing the db helps you with backup. `mongodump` does not work on the datafiles any more (and selected the database and collections automagically anyway) and copying the files of one database only away does not work, too (namespaces, oplog). The remaining option is to make a filesystem snapshot, for which you do not need the databases identified. – Markus W Mahlberg Dec 10 '15 at 10:50
  • Side: It is not the best idea to run a database on a non-LTS release – there is a reason why MongoDB Inc. offers Ubuntu Packages only for Pangolin and Tahr. – Markus W Mahlberg Dec 10 '15 at 10:54

0 Answers0