Jira (PDB-4342) Add a node lifetime timestamp, affected by configure expiration

25 views
Skip to first unread message

Rob Browning (JIRA)

unread,
Jul 25, 2019, 12:00:16 PM7/25/19
to puppe...@googlegroups.com
Rob Browning updated an issue
 
PuppetDB / Improvement PDB-4342
Add a node lifetime timestamp, affected by configure expiration
Change By: Rob Browning
Summary: Add a node lifetime " timestamp " , affected by configure expiration
Add Comment Add Comment
 
This message was sent by Atlassian JIRA (v7.7.1#77002-sha1:e75ca93)
Atlassian logo

Rob Browning (JIRA)

unread,
Jul 25, 2019, 12:04:04 PM7/25/19
to puppe...@googlegroups.com
Rob Browning updated an issue
Add a node "timestamp" (perhaps in the certnames table) that behaves will behave like the other lifetime-related timestamps, i.e. the factsets/reports timestamps

Extend
, and adjust the configure expiration command to allow affect it when the " node expires " as a value in the field includes " expire node " .  (Not sure yet whether configure expiration should initially allow more than one field in expires , i.e. { " node facts " : <boolean> ...} and have the command behave as it does for facts, but


Support a way for users to explicitly refresh the timestamp used for
" node TTL.

This can be accomplished currently by repeatedly sending *configure expiration* commands
" -- not critical for a node with *expire now . facts = true* (see linked ticket for more detail )

We should make this more explicit (maybe adding a new command) This ticket only needs to cover the schema/migration and look into command adjustment.  Related tickets will cover the possibility of allowing bulk updates given a list of certnames remaining work .

Rob Browning (JIRA)

unread,
Jul 25, 2019, 12:07:03 PM7/25/19
to puppe...@googlegroups.com
Rob Browning commented on Improvement PDB-4342
 
Re: Add a node lifetime timestamp, affected by configure expiration

Excellent point. After discussion, a good bit of it with Zak while working on the other configure expiration related work, adding a real node timestamp seems like the best way forward, so I've adjusted the ticket accordingly.

Rob Browning (JIRA)

unread,
Jul 25, 2019, 12:08:03 PM7/25/19
to puppe...@googlegroups.com
Rob Browning commented on Improvement PDB-4342

Also adjusting the epic (and it's tickets) to reflect the amended plan.

Rob Browning (JIRA)

unread,
Jul 25, 2019, 12:15:03 PM7/25/19
to puppe...@googlegroups.com
Rob Browning updated an issue
Change By: Rob Browning
Add a node "timestamp" (perhaps in the certnames table) that will behave like the other lifetime-related timestamps, i.e. the factsets/reports timestamps, and adjust the configure expiration command to affect it when the "expires" field includes "node".  (Not sure yet whether configure expiration should initially allow more than one field in expires, i.e. "facts" and "node" -- not critical for now.)

This ticket only needs to cover the schema/migration
and , command adjustment , and attendant docs (e . g. the configure expiration wire format, and possibly information about the new node timestamp).   Related tickets will cover the remaining work.

Rob Browning (JIRA)

unread,
Jul 25, 2019, 12:18:03 PM7/25/19
to puppe...@googlegroups.com
Rob Browning updated an issue
Add a node "timestamp" (perhaps in the certnames table) that will behave like the other lifetime-related timestamps, i.e. the factsets/reports timestamps, and adjust the configure expiration command to affect it when the "expires" field includes "node". , i.e. {"certname": ... "expire": {"node": <boolean>} ...}.   (Not sure yet whether configure expiration should initially allow more than one field in expires, i.e. "facts" and "node" -- not critical for now.) .  The configure expiration command should behave like it does for facts, maybe reactivating the node, etc.

This ticket only needs to cover the schema/migration, command adjustment, and attendant docs (e.g. the configure expiration wire format, and possibly information about the new node timestamp).  Related tickets will cover the remaining work.
Reply all
Reply to author
Forward
0 new messages