Long running Discovery Schedules
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
‎05-07-2018 08:24 AM
We recently upgraded from Istanbul to Kingston version, the issue is with the Discovery Schedules. It is taking more time to complete when compared to our previous version. Can someone help me in knowing what makes it to run such long time when compared it to Istanbul version.
- Labels:
-
Discovery
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
‎05-07-2018 08:58 AM
Well, first, let's identify the slow probes.
Go to the Discovery Status of a completed slow discovery, click on the ecc queue tab, and sort on "created" from newest to oldest.
I bet you see a few probes that finish a lot later than the crowd.
What are the names of those probes?
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
‎05-17-2018 07:08 AM
Hey Sorry for late reply, it's SNMP probe taking more than 1hour.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
‎05-17-2018 01:09 PM
Oh, yeah. SNMP is definitely touchy by nature.
Knobs to consider fiddling with include:
- turning off probes you don't need, e.g. snmp - routing probes, which bring in huge results whose contents you may not care about
- tuning the timeout parameter
- use_getbulk (for tables)
- use_getscalar (for scalars)
- generally squinting at packet captures and trying to figure out where the time is going
I'm afraid I still need help every time I hit one of these issues with SNMP.
You might want to open an incident if you can't sort it out yourself within an hour or two.
I'm afraid empathy is most of what I can offer there. Let us know what you figure out?
- Tim.