Last Check-in Still updating even after failed run


#1

We just upgraded to 11.4.0 from 10.24.0, so I am not sure if this is a feature
of 11,

When chef-client runs, it will update the “Last check-in” column for the node’s
status in our hosted Chef even if the run failed.

In 10.24.0 this was not the case. If the run failed, the check-in was not
updated, which was how we knew the node was having problems.

Is there a way to turn this back on, or is there supposed to be another I
should find out how the node chef-client run failed??


#2

Yes , this is a chef 11 feature. It runs delayed notification even after
failures. It was highly debated. I had filed a ticket to make this at least
configurable , but it looks like even that won’t happen in chef 11. Chef 12
might have something on that line…
On Aug 7, 2013 9:25 AM, “ScottMarshall” scott.marshall@johnmuirhealth.com
wrote:

We just upgraded to 11.4.0 from 10.24.0, so I am not sure if this is a
feature
of 11,

When chef-client runs, it will update the “Last check-in” column for the
node’s
status in our hosted Chef even if the run failed.

In 10.24.0 this was not the case. If the run failed, the check-in was not
updated, which was how we knew the node was having problems.

Is there a way to turn this back on, or is there supposed to be another I
should find out how the node chef-client run failed??


#3

On Wednesday, August 7, 2013 at 9:34 AM, Ranjib Dey wrote:

Yes , this is a chef 11 feature. It runs delayed notification even after failures. It was highly debated. I had filed a ticket to make this at least configurable , but it looks like even that won’t happen in chef 11. Chef 12 might have something on that line…
On Aug 7, 2013 9:25 AM, “ScottMarshall” <scott.marshall@johnmuirhealth.com (mailto:scott.marshall@johnmuirhealth.com)> wrote:

We just upgraded to 11.4.0 from 10.24.0, so I am not sure if this is a feature
of 11,

When chef-client runs, it will update the “Last check-in” column for the node’s
status in our hosted Chef even if the run failed.

In 10.24.0 this was not the case. If the run failed, the check-in was not
updated, which was how we knew the node was having problems.

Is there a way to turn this back on, or is there supposed to be another I
should find out how the node chef-client run failed??
Chef does run delayed notifications even if chef failed as of Chef 11. However, this should not result in chef-client doing a node save after a failed chef-client, and in my own testing I cannot reproduce this behavior.

What you may be seeing is the effect of having a node.save call somewhere in your recipes that executes before the run fails, or a node.save call in a resource that gets a delayed notification. Try running chef-client with -l debug to see if you can track down the source of the issue.


Daniel DeLeo