Ask Your Question
1

How should Member Node operators inform the DataONE community of downtime of a repository?

asked 2013-03-22 18:27:58 -0500

cjones gravatar image

There are times when repository systems need to go offline for routine maintenance, or to deal with unplanned outages. In what way should I tell the DataONE community that my Node is unavailable temporarily? Is there a mailing list? Is there a formal Node setting to let the system know programmatically?

edit retag flag offensive close merge delete

2 answers

Sort by ยป oldest newest most voted
1

answered 2013-05-01 09:06:48 -0500

updated 2013-05-01 10:53:51 -0500

For downtime, node operators should take the following actions:

  1. For scheduled downtime:

    • send an email to 'operators@dataone.org' to provide as much notice as practical. If the scheduled downtime is more than a few days out, sending a follow-up email to 'operators@dataone.org' the day before the downtime is appreciated.
    • Just before the outage starts, use the Member Node software to inform the coordinating nodes that the MN is down. This is done by by calling the updateNodeCapabilites() method and marking the node state as 'down'. The exact details of how to call this method depend on how the Member Node has been implemented. See the question about updating node information for more information about how to do this update.
    • After the MN is back up, update the node state to 'up' using the updateNodeCapabilities call as above and send a follow-up email to 'operators@dataone.org' to say that the MN is back up.
  2. For unscheduled outages:

    • Send an email to 'operators@dataone.org' to say that the MN is down. Provide as much information as you can about the expected duration of the outage. Please indicate whether you're able to update the node status, as described above, and provide a contact phone number. If you're not able to update the node status, we may do so on your behalf, depending on the anticipated length of the outage.
    • Provide updates via email to 'operators@dataone.org' as practical during the outage, with updates on when return to service is expected.
    • When the node is back on-line, notify DataONE via an email to 'operators@dataone.org' and use a call to updateNodeCapabilities, as noted under scheduled outages, to make sure the node status is set to 'up'.
edit flag offensive delete link more
0

answered 2013-03-27 13:25:27 -0500

Node operators should perform two actions: first, announce to the 'operators@dataone.org' mailing list that the node's services will be inaccessible, and summarize the duration and reason for the outage; second, when the node services are to be taken offline, update the node's description document to mark the node state as 'down'. This is accomplished by calling the updateNodeCapabilites() service on the Coordinating Nodes. Once the service outage is over, update the node to the 'up' status using this same service, and announce that the services have been restored to the operators@dataone.org email list.

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

[hide preview]

Question Tools

1 follower

Stats

Asked: 2013-03-22 18:27:58 -0500

Seen: 162 times

Last updated: May 01 '13