Hi All, In the scenario where the ovsdb has already updated the database , but the application that has registered for monitoring the schema has failed to process DB updates ( due to internal errors ) , we have to roll back the database changes.
Is there a mechanism available in ovsdb-server to roll back to a certain previous commit point. I would greatly appreciate any response on this as it would help us in making right design decisions. Thanks and regards, Rishi Raj ---------- Forwarded message ---------- From: RishiRaj Maulick <rishi.raj2...@gmail.com> Date: Mon, Sep 21, 2015 at 5:01 PM Subject: Query with respect to DB commit even when vsctl failed To: disc...@openvswitch.org, dev@openvswitch.org Hi , I have been using OVS recently and I noticed the following - 1) I did not insert the openvswitch kernel module 2) Using “ovs-vsctl add-br br0” command , I tried to add a bridge – which failed for obvious reasons. There is an error log that is displayed on ovs-vsctl screen as well. 3) However there are entries that are left in the ovs database – “ovsdb-client dump”. I can see addition to the bridge , port , interface and openvswitch tables – all entries pertaining to the bridge “br0”. Ideally an application has to delete the bridge “br0” from the DB after hitting an error like this. I see that the ovs-vsctl can detect these errors through checks for the field in the interface table entry "ofport”. The question is how can a remote management interface for any proprietary schemas get to know about application processing failures . Thanks and regards, Rishi Raj Maulick _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
