[4.0] Decrease the amount of node saves #102
Conversation
Rebase of PR crowbar#80 Only save the node when attribute(s) have changed Do not fetch monitor secret if we're the master
| dirty = true | ||
| end | ||
| # if journal_device is nil, this will still work as expected | ||
| if node["ceph"]["osd_devices"][index]["journal"] != journal_device |
There was a problem hiding this comment.
Metrics/BlockNesting: Avoid more than 3 levels of block nesting. (https://github.com/bbatsov/ruby-style-guide#three-is-the-number-thou-shalt-count)
| end | ||
| node.set["ceph"]["osd_devices"][index]["status"] = "deployed" | ||
| node.set["ceph"]["osd_devices"][index]["journal"] = journal_device unless journal_device.nil? | ||
| if node["ceph"]["osd_devices"][index]["status"] != "deployed" |
There was a problem hiding this comment.
Metrics/BlockNesting: Avoid more than 3 levels of block nesting. (https://github.com/bbatsov/ruby-style-guide#three-is-the-number-thou-shalt-count)
|
|
||
| node.set["ceph"]["osd"]["use_ssd_for_journal"] = false unless has_ssds && has_hdds | ||
| use_ssd_for_journal = has_ssds && has_hdds | ||
| if node["ceph"]["osd"]["use_ssd_for_journal"] != use_ssd_for_journal |
There was a problem hiding this comment.
Metrics/BlockNesting: Avoid more than 3 levels of block nesting. (https://github.com/bbatsov/ruby-style-guide#three-is-the-number-thou-shalt-count)
| if !node["ceph"]["osd"]["journal_devices"].empty? | ||
| node.set["ceph"]["osd"]["use_ssd_for_journal"] = false | ||
| # explicit comparison because we don't want a condition that uses nil | ||
| if node["ceph"]["osd"]["use_ssd_for_journal"] != false |
There was a problem hiding this comment.
Metrics/BlockNesting: Avoid more than 3 levels of block nesting. (https://github.com/bbatsov/ruby-style-guide#three-is-the-number-thou-shalt-count)
| device["device"] = d.name | ||
| node.set["ceph"]["osd_devices"].push(device) | ||
| node.save | ||
| dirty = true |
There was a problem hiding this comment.
We were saving here explicitly for a reason, probably becasue we are using this values ["ceph"]["osd_devices"] immediately below.
Are we sure this is still working as before? I would say yes, as its only used in this recipe, but would like confirmation that this has been tested properly and the devices are being picked up in the recipe afterwards without saving the node.
There was a problem hiding this comment.
Yes, the array of devices is enumerable (i.e. usable) immediately below even after moving node.save to the end of the file. The node.save is only needed to persist the values for subsequent executions of the cookbook.
Itxaka
left a comment
Itxaka
left a comment
There was a problem hiding this comment.
Looks good, just a question and a nit
Itxaka
left a comment
Itxaka
left a comment
There was a problem hiding this comment.
Looks good, just a question and a nit
Itxaka
changed the title
| dirty = false | ||
|
|
||
| node.set["ceph"]["osd_devices"] ||= [] | ||
| node.set["ceph"]["osd_devices"] = [] if node["ceph"]["osd_devices"].nil? |
Fix multiline comment Fix array initialization (was mistakenly initializing as a hash)
scottwulf
force-pushed
the
decrease-node-save
branch
from
55455d5 to
ad1fbcc
Compare
4 participants
Rebase of PR #80
Only save the node when attribute(s) have changed
Do not fetch monitor secret if we're the master