Ask Your Question

What is the exact procedure to be done if one of hard drives failed in storage node of swift?

asked 2014-09-02 03:47:36 -0500

ebyenjoys gravatar image

updated 2014-09-03 22:51:44 -0500


I do have object storage deployment with three replicas. The three replicas was configured as three hard drives in my object storage node. I am not aware of the procedures to be done if one of my hard drives got damaged. It would be a great help for me if some one could assist me to teach how to remove the damaged hard drive and replace it with a new device.

edit retag flag offensive close merge delete

1 answer

Sort by ยป oldest newest most voted

answered 2014-09-03 21:38:31 -0500

zaitcev gravatar image

updated 2014-09-03 21:39:17 -0500

Unmount and remove the old drive. Install the new drive, run mkfs.xfs as usual and mount it like the old drive under /srv/node. Make sure it mounts on boot like the old drive.

At this point there's going to be a spike of replication load, since your cluster is so small. If that's undesirable, you can do the usual dance with swift-ring-builder and starting with close-to-zero weight (don't forget to push the ring). See:

If the load spike is not a big concern, I would not even bother with any changes to the ring at all.

edit flag offensive delete link more


Thanks a lot zaitcev.

Let me introduce one scenario. I do have sdb,sdc and sdd hard drives. Even if one of my hard drive fails the other two drives could deliver the files.Suppose the sdd hard drive got damaged. If I introduce a new hard drive to replace the damaged sdd, will it be labeled as sdd?

ebyenjoys gravatar imageebyenjoys ( 2014-09-06 14:20:35 -0500 )edit

It would be great that if the new hard drive could work exactly as like sdd,so that i don't want to make any unnecessary changes in the Ring.

ebyenjoys gravatar imageebyenjoys ( 2014-09-06 14:25:48 -0500 )edit

Get to know Ask OpenStack

Resources for moderators

Question Tools

1 follower


Asked: 2014-09-02 03:47:36 -0500

Seen: 466 times

Last updated: Sep 03 '14