Update the code to use scontrol in place of slurm APIs to drain the #74
Conversation
mithun-pensando
force-pushed
the
redfishmh
branch
2 times, most recently
from
3b048a6 to
478801c
Compare
| @@ -219,15 +220,34 @@ func (s *Server) processRequest(AppConfig Config, conn net.Conn, req *http.Reque | |||
| log.Printf("Origin Of Condition: %s", originOfCondition) | |||
| for _, triggerEvent := range AppConfig.TriggerEvents { | |||
| if severity == triggerEvent.Severity { | |||
| log.Printf("Matched Trigger Event: %s with action %s", triggerEvent.Severity, triggerEvent.Action) | |||
| if triggerEvent.Message != "" { | |||
There was a problem hiding this comment.
So, the action will only be triggered if the severity message matches as well?
There was a problem hiding this comment.
If the user doesn't want to match on the 'message' field, it can be left empty. I will add it to the comments in .env
| TRIGGER_EVENTS="[\ | ||
| {\"Severity\":\"Fatal\",\"Action\":\"DrainNode\"},\ | ||
| {\"Severity\":\"Critical\",\"Action\":\"DrainNode\"} | ||
| {\"Severity\":\"Critical\",\"Message\":\"message 1|This is a critical test event\",\"Action\":\"DrainNode\", \"DrainReasonPrefix\":\"RebootNeeded\"},\ |
There was a problem hiding this comment.
Is an action necessary if it's just going to be a drain anyway?
Could you also add a comment for these fields as well to make it clear?
There was a problem hiding this comment.
I left it as is in case we need to use it for some other action in the future. Will update the comments
redfish-exporter/slurm/slurm.go
Outdated
| @@ -160,6 +201,183 @@ func (c *Client) GetNodes() ([]string, error) { | |||
| return nodes, nil | |||
| } | |||
|
|
|||
| func (c *Client) GetNodeReason(nodeName string) (string, error) { | |||
There was a problem hiding this comment.
Is this code needed anymore?
There was a problem hiding this comment.
I left in the functions which use API for future use. I will update the name of this function
| "log" | ||
| "strings" | ||
|
|
||
| "github.com/nod-ai/ADA/redfish-exporter/metrics" | ||
| ) | ||
|
|
||
| const ( | ||
| Drain = "DrainNode" | ||
| Drain = "DrainNode" | ||
| ExlcudeReasonSet = "DRAIN_EXCLUDE_REASON_SET" |
There was a problem hiding this comment.
Does the user provide this? Is this different than the exclude reason from .env?
There was a problem hiding this comment.
This is for internal logging
node to make it slurm version independent
mithun-pensando
force-pushed
the
redfishmh
branch
from
478801c to
034e2ad
Compare
| @@ -160,6 +201,183 @@ func (c *Client) GetNodes() ([]string, error) { | |||
| return nodes, nil | |||
| } | |||
|
|
|||
| func (c *Client) GetNodeReasonWithAPI(nodeName string) (string, error) { | |||
There was a problem hiding this comment.
Is this needed now that we are moving everything to scontrol?
There was a problem hiding this comment.
I had tested this API path also, left it in if we move to API later
node to make it slurm version independent