second to know how many appliances you have in the GMS.
Third, GIVE IT MORE.
In my experience, he always wants more, I have 2 GMS v9.2 an 1 GMSv9.3, one of the GMS9.2 has 45 appliances, (the other 9.2 only 10 appliances) (all under VMWare 6.5) and I have assigned it, 24G to RAM and 4 CPU cores and it always eats everything up.
But it is not something that you have to worry about if it does its job. (If the high CPU suo seems strange to me) The problem is when something goes wrong, and it stops summarizing, and starts not processing the syslogs.
We had 2 GMS v9.2 with 30G to RAM and 16 CPU cores for each. Both has 22 appliances. As you can see, we have a lot of resources assigned to GMS. But even so, we had this problem of total resource consumption. This issue was only resolved when we upgraded to GMSv9.3. But we don't know if it was the update or the fact that we restarted the appliances that solved the problem.
So, is it normal that, over time, the GMS gradually increases the resources consumption to the point of exhaustion?
I can only talk about my experience, and yes, but in memory, not in CPU, for some reason that I do not know the memory, it runs out, and the machine begins to not do or not be able to do its tasks, and I am forced to restart it, especially to this one of 45 nodes. the other 10, very from time to time.
Once rebooted, it has huge amounts of Syslogs, raw for 3 or 5 days, and then yes, the CPU starts to process everything that it has not been able to do the days before. making intensive use of it.
This is the graph of 5 days, the GMS 9.2 of 10 nodes.
In red, the normal memory consumption by the Syslogs processing (which I think is not normal, every 15 min it processes the Syslogs that arrive and it does not take more than 5 min to process them all, but it is a linear memory consumption)
In blue when it processes the reports, and then drops sharply once finished. The difference of falling in pink, sometimes does not occur, and ends, collapsing due to lack of RAM, As I say, are my observations, in my environment, and as the number of appliances and their specific behaviors. It does not have to match other systems.
For me, v9.2 is going very well, compared to other previous versions, similar to 8.2, which also worked very well.
I hope that since you have migrated to 9.3, you share your feelings and observations with this version, which seems to be the last.
Answers
The first thing would be to know if you meet the minimum requirements, for GMS,
https://www.sonicwall.com/techdocs/pdf/global-management-system-9-2-getting-started-guide.pdf
second to know how many appliances you have in the GMS.
Third, GIVE IT MORE.
In my experience, he always wants more, I have 2 GMS v9.2 an 1 GMSv9.3, one of the GMS9.2 has 45 appliances, (the other 9.2 only 10 appliances) (all under VMWare 6.5) and I have assigned it, 24G to RAM and 4 CPU cores and it always eats everything up.
But it is not something that you have to worry about if it does its job. (If the high CPU suo seems strange to me) The problem is when something goes wrong, and it stops summarizing, and starts not processing the syslogs.
The capture GMS9.2 45 appliaces. last 5 days.
We had 2 GMS v9.2 with 30G to RAM and 16 CPU cores for each. Both has 22 appliances. As you can see, we have a lot of resources assigned to GMS. But even so, we had this problem of total resource consumption. This issue was only resolved when we upgraded to GMSv9.3. But we don't know if it was the update or the fact that we restarted the appliances that solved the problem.
So, is it normal that, over time, the GMS gradually increases the resources consumption to the point of exhaustion?
Many thanks for the reply @Chusito.
I can only talk about my experience, and yes, but in memory, not in CPU, for some reason that I do not know the memory, it runs out, and the machine begins to not do or not be able to do its tasks, and I am forced to restart it, especially to this one of 45 nodes. the other 10, very from time to time.
Once rebooted, it has huge amounts of Syslogs, raw for 3 or 5 days, and then yes, the CPU starts to process everything that it has not been able to do the days before. making intensive use of it.
This is the graph of 5 days, the GMS 9.2 of 10 nodes.
In red, the normal memory consumption by the Syslogs processing (which I think is not normal, every 15 min it processes the Syslogs that arrive and it does not take more than 5 min to process them all, but it is a linear memory consumption)
In blue when it processes the reports, and then drops sharply once finished. The difference of falling in pink, sometimes does not occur, and ends, collapsing due to lack of RAM, As I say, are my observations, in my environment, and as the number of appliances and their specific behaviors. It does not have to match other systems.
For me, v9.2 is going very well, compared to other previous versions, similar to 8.2, which also worked very well.
I hope that since you have migrated to 9.3, you share your feelings and observations with this version, which seems to be the last.
Greetings.