Skip to content

Commit

Permalink
[FAB-6675] Document sqlite db locked error
Browse files Browse the repository at this point in the history
The "database locked" error is expected when sending
multiple parallel requests to Fabric CA server cluster
that is connected to a sqlite database. This configuration
is not supported and valid. Added a section to the
troubleshooting section of the user's guide to document
this.

Change-Id: Ic044b459d228fd4b4d5ab59de013ad833344734e
Signed-off-by: Anil Ambati <aambati@us.ibm.com>
  • Loading branch information
Anil Ambati committed Oct 26, 2017
1 parent 2886abd commit 37cbb14
Showing 1 changed file with 8 additions and 0 deletions.
8 changes: 8 additions & 0 deletions docs/source/users-guide.rst
Original file line number Diff line number Diff line change
Expand Up @@ -1479,5 +1479,13 @@ Troubleshooting
To resolve this error, you must enroll again by repeating step 'a'. This will issue a new ECert
which will be stored in the current database.

3. When sending multiple parallel requests to a Fabric CA Server cluster that uses shared sqlite3 databases,
the server occasionally returns a 'database locked' error. This is most probably because the database
transaction timed out while waiting for database lock (held by another cluster member) to be released.
This is an invalid configuration because sqlite is an embedded database, which means the Fabric CA server
cluster must share the same file via a shared file system, which introduces a SPoF (single point of failure),
which contradicts the purpose of cluster topology. The best practice is to use either Postgres or MySQL
databases in a cluster topology.

.. Licensed under Creative Commons Attribution 4.0 International License
https://creativecommons.org/licenses/by/4.0/

0 comments on commit 37cbb14

Please sign in to comment.