How can we update inventory of 4000 computers once every day?

You can create a policy, in the general tab make sure you have the option, "Once per day" selected. Then add the "Maintenance" payload and select "Update inventory". Be sure to scope it to the proper computers or group, building etc...

as a side note, this can use considerable network bandwidth. We update once per week to have a smaller affect on the network.

So out of curiosity, what are your call home times? Inventory updates versus check-ins? Are you still seeing check-ins, but not inventory updates?

@chenhao2018 What is limiting you to 1200 computers a day? We don't have anywhere near that number but I'm intrigued by what this limit you're running into is. How are you assigning the policy to the machines? All computers, static group, smart group, etc?

A default Jamf Pro setup includes a policy that captures inventory once per day scoped to all managed devices. Unless you removed that, it should be there, unless it's something new that Jamf isn't including that in newer releases. If you're creating a new policy, there should be nothing limiting you to only being able to choose a random 1200 machines vs all of them.

Or did you mean you're only getting about 1200 of them updating their inventory? If so, that might just be that the other Macs aren't active or have yet to check in for inventory update. They don't all do it at the same time. It's random on purpose.

I did just like clegger06 said, but when i see the policy log, there are only 1200+ macs updated per day. I am thinking if it is because of the jss low capability.

@jhuls all computers scoped. I mean we have 4000+, but one day only 1000+ updated.

I am guessing thats probably pretty normal, computers would only update if they were powered on more than 24 hours after the previous update so it may take several days to capture all machines depending on the usage pattern.

@Look Is that possible we create two same update policies with different scopes. We need update 4000+ one day.

@chenhao2018 Can you screen shot your policy’s general, maintenance and scope sections? Make sure to include the triggers and execution frequency

I have almost 4000 macOS devices and any given day about 3000 are online given day. They all update inventory just fine with the right policy.

the setting is right？@cubandave

@chenhao2018 Looks ok to me. I guess I'm not sure what the main issue is with this? Are you just hoping that all Macs would submit inventory at the same time each day, or is there some other issue with how it's working?

Are all of your computers online during the day? Are they all on a network that can reach the JSS? Are they checking in?

If you take a look at /var/log/jamf.log on some of the machines that are failing to update inventory, it might give you a clue as to why.

Your setup looks fine, and there should be no issue with JSS capcacity. Even a single JSS should be able to handle 4000 clients reasonably well, assuming that your server hardware is decent.

@mm2270 I am very interested in the inner principle of jss inventory update. What the order of clients? for example, we have 4000 macs, and we set the check-in time 15mins. if we have 1000 macs has been checked in, and wait for the updating one by one. then i want to update another mac which is not in the 1000 queue, is that this mac still have to wait? update one by one is not good!

@mjhersh Actually those macs have to connect to company's inner internet for being managed. if they are in home, without vpn, they can not get updated. The problem is how can we know which macs are out of connected?

Posted: Today at 12:03 PM by chenhao2018 @mjhersh Actually those macs have to connect to company's inner internet for being managed. if they are in home, without vpn, they can not get updated. The problem is how can we know which macs are out of connected?

In order for the macs that are at home to be able to check in you will need to have an internet facing JSS or DMZ JSS. You can use split DNS to resolve both.

Check out the clustering in the Jamf Pro Admin Guide.

@mm2270 I am very interested in the inner principle of jss inventory update. What the order of clients? for example, we have 4000 macs, and we set the check-in time 15mins. if we have 1000 macs has been checked in, and wait for the updating one by one. then i want to update another mac which is not in the 1000 queue, is that this mac still have to wait? update one by one is not good!

The JSS does not control when the inventory happens. There is no queue in the JSS that would control this. It is controlled by the jamf recon binary on the client. So if a client last submitted an inventory at 9:01am yesterday it will not submit another inventory for 24 hours from that time or the next check in after that 24 hours has passed. The inventory won't just happen "once per day", so the execution frequency in the policy is somewhat misleading. Once per day does not mean once per calendar day, it means once per 24 hour period. Same for the once per week, once per month frequencies. If the client is off the network, powered off, etc it cannot submit an inventory and should during the next check-in once it is powered/returns to the network after the 24 hours has passed since the last inventory. So the inventory submission will always be pretty much random across all of your clients since scheduled check-ins are also controlled by the binary on the client and not the JSS.

The only way you could maybe get true once per day inventory is to trigger it from the Jamf Remote app daily or create some kind of custom script/launchd that would trigger the jamf recon binary on the client each day at a specific time.

@scottlep good explaination. some kind of custom script/launchd. l know some sh, but how to use crontab, I see someone said crontab doesn't work any more in macOS. how to use launchd. Do you have any documents about it?