Ask Your Question
1

swift container-sync not working

asked 2013-05-15 12:38:19 -0500

simplidrive gravatar image

Hi,

We are running two separate clusters of swift 1.5.0 with swauth 1.0.6.

We are trying to set up container sync between accounts on both the cluster, but getting below error on container server on one of the clusters :

May 15 16:35:42 Zone4-S001 container-sync Unauth 'AUTH_25d97b67-134c-49a5-bf61-194047787d1f/container1' => 'https://simplidrivenoi.netmagicsolutions.com/v1/AUTH_fcecf343-f35f-423a-bfae-e8200d3f6293/container2'

We have configured below parameters on all the storage nodes (no configuration done on proxy servers) :

1) in container-server.conf file : under [DEFAULT] section allowed_sync_hosts = http://127.0.0.1,180.179.199.181,simplidrivenoi.netmagicsolutions.com (127.0.0.1,180.179.199.181,simplidrive...)

under [container-sync] section : log_name = container-sync log_level = DEBUG

Is there anything we are missing here ?

Regards,

Viral Patadiya

edit retag flag offensive close merge delete

4 answers

Sort by ยป oldest newest most voted
0

answered 2013-05-31 20:39:59 -0500

dmitry-pha gravatar image

Just solved the same problem.

In proxy-server.conf [filter:swauth] section I had the hostname specified in allowed_sync_hosts, while the system was expecting IP address

edit flag offensive delete link more
0

answered 2013-11-01 09:20:04 -0500

simplidrive gravatar image

Thanks dimm, that solved my question.

edit flag offensive delete link more
0

answered 2013-05-17 07:48:44 -0500

notmyname gravatar image

That looks correct for configuration, but the log line you are seeing is from swauth. Looking at the code, it's hard to tell exactly what's causing it (it's a catch-all log line). One possibility is that your clusters clocks are out of sync.

edit flag offensive delete link more
0

answered 2013-05-21 09:53:13 -0500

simplidrive gravatar image

Hi,

The clocks are already in sync. Even the sync is not happening between two accounts in the same cluster.

couple of queries :

  1. Is it because we have a LB with ssl in front of proxies ?
  2. Is it mandatory to have storage nodes on public IPs? We just have public IPs for the sake of management, otherwise the storage segment is on private IPs only.
  3. have tried syncing two containers from two different account on the same cluster, even that is not happening.

Any more ideas ?

Regards,

Viral

edit flag offensive delete link more

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

Get to know Ask OpenStack

Resources for moderators

Question Tools

1 follower

Stats

Asked: 2013-05-15 12:38:19 -0500

Seen: 101 times

Last updated: Nov 01 '13