ipdevpoll jobs that are no longer supposed to run for a device as flagged as "overdue"
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Network Administration Visualized |
Fix Released
|
Medium
|
Morten Brekkevold |
Bug Description
When ipdevpoll actively decides not to run a particular job for a device, no further references to that job are logged for this device.
This is what happens with the `ip2mac` job if, for example, a device's category is changed from GW/GSW to SW. The `ip2mac` job runs the `arp` plugin every 30 minutes, but this plugin will only want to run for routers.
At this point, no more `ip2mac` jobs are logged for this device - yet tools like ipdevinfo and WatchDog will claim that this job has become overdue for this device. ipdevinfo and WatchDog cannot know that ipdevpoll has decided that the job is no longer relevant for this device.
ipdevpoll still will check every 30 minutes whether to run the `ip2mac` job for an SW device, so probably the best way to solve this quandary is to also log these checks as "not run" jobs (i.e a tertiary status value, as opposed to todays "success/failure" boolean status).
Changed in nav: | |
status: | Fix Committed → Fix Released |
Changed in nav: | |
milestone: | 4.2.4 → 4.2.5 |
fix here: https:/ /nav.uninett. no/hg/stable/ rev/250fdc2acc5 4