Jump to content

Problems with MDM and High Server CPU %


HSW
 Share

Recommended Posts

Hi we have some problems with our ERA.

Since last week the cpu % of the Server goes from 5 to 90% every 5s.

I checked the running services and the MDM is the problem.

 

Also some clients report some strange problems. One send one as attachment.

The status of the clients are red or orange from some past status. The server dont get the new status, but i see the correct last connection.

We try to restart the server and also the clients. On client side all is fine.

ERA and MDM have the latest version.

Clients have also the newest version.

 

There are some known problems? Any tips?

 

android1.png

android2.png

android3.png

era1.png

era2.png

era3.png

Link to comment
Share on other sites

Update:

I have restore the DB, after this clients status is ok.

High CPU Usage of MDM ist still to high. Any Help?

Link to comment
Share on other sites

  • 2 weeks later...

We still have the high cpu usage with the MDM Modul. Any fixes here? Its not a nice status for a Server :(

Link to comment
Share on other sites

  • ESET Staff

Hello,

could you check the MDM database, the row counts in the various tables? Especially the tbl_ma_dataminer_deltalog table, how many rows does it contain, whether hundreds, thousands, tens of thousands.

Thank you.

Link to comment
Share on other sites

  • ESET Staff

That seems like a failed cleanup. Please try executing this SQL query in the MDM DB, then check if the number of rows has decreased substantially:

DELETE FROM tbl_ma_dataminer_deltalog WHERE KeyFrameNo < -10

Link to comment
Share on other sites

  • ESET Staff

That's a much better number. The CPU usage should gradually go down.

Can you check if there are any more tables in the DB that have more than a 5000 rows in them? The following SQL query produces a list of tables and approximate row counts:

show table status

What is your exact MDM version?

Link to comment
Share on other sites

  • ESET Staff

There is a new version of Mobile DeviceConnector in the final stages of QA, it should be available within a few days. Please upgrade to it, once available, then re-run the DELETE sql query, just to be sure. If it still has a high CPU usage, please do contact the support, they'll ask for more information and will try to help you individually.

Link to comment
Share on other sites

Thx for your help, sorry i overread your last post, the deltalog is again big :(

 

This tables have more than 5K:

EnrollmentTokens 30K

tbl_ma_dataminer_deltalog 660K

tbl_static_objects 13K

tbl_static_objects_data 13K

 

I can wait for the new update, if i should do every think else write it :)

EDIT: MDM Version see first Post: 6.5.391.0

 

Edited by HSW
Link to comment
Share on other sites

i cleaned the deltalog yesterday again and till now the entrys stay small. How can i cleanup the other tables too?

I also improve the resources to 4 CPUs und 10 GB Ram (only temporary till the problem is gone, 100% cpu is now gone max is 70%)

Edited by HSW
Link to comment
Share on other sites

  • ESET Staff

The deltalogs growth will be fixed by the new release, the other tables are not performance-critical and shouldn't be cleaned up in any way.

Note that the MDC is expected to have a high base CPU usage, and only very small increments in usage for each additional device managed. That's one of the main reasons why we recommend to not instal the MDC on the same machine as the ERA server. As long as the CPU is not constantly at 100 %, there's no reason to worry, really.

Until you have the new version installed, it is very likely that no tasks and no logs will be delivered between ERA and MDC, so you should upgrade as soon as it's possible.

Link to comment
Share on other sites

18 hours ago, madmaxoft said:

Note that the MDC is expected to have a high base CPU usage, and only very small increments in usage for each additional device managed. That's one of the main reasons why we recommend to not instal the MDC on the same machine as the ERA server. As long as the CPU is not constantly at 100 %, there's no reason to worry, really.

Thats NOT nice to read, why you dont write such essential information in the docu in red? Any information of the needed hardware resources? https://support.eset.com/kb5773/

Please give me some infos and i think about a change of the mdm to a standalone system.

You say its normal with the high cpu ussage, i must say you: no this wasnt in the past.

You say there is no problem with the high ussage -> not realy normal actual we have only 16 devices. What if we have 100? Need a 8 core cluster?

Link to comment
Share on other sites

  • ESET Staff

Sorry for the late answer.

The hotfix has just been released.

I thought we had instructed the documentation team to put the info in the docs, but I guess we weren't thorough enough. We'll fix that.

If the CPU usage is higher than it used to be for you, then it's a sign there's something wrong, that's for sure. There shouldn't be any increase. You should contact the support in this case, as they will request more details that are not wise to share on a public forum. Anyway, as for the CPU usage, MDC is built in such a way that the CPU usage doesn't depend on the number of managed devices. So it may seem a bit too much for a few devices, but it will stay more or less the same even if you manage hundreds or thousands of devices. Of course nobody is perfect, so there might be a hidden edge case, a bug that you might have come across, so we would like to know and fix it.

 

Having MDC on a separate system has a few benefits, so you should consider it anyway. The MDC machine must be visible from public internet, while ERA server could be hidden behind firewalls / NAT / ... . The MDC needs a hostname to work (so that the phones have a fixed address to contact), using the ERA hostname (or even worse, an IP address) binds your hands unnecessarily. Performance would be another point.

Link to comment
Share on other sites

Hi i updated the mdc und clean up the tbl_ma_dataminer_deltalog again + reboot the machine.

Actualy it looks better withe the cpu usage. I test with my device and now status updates from the mobile gets updated.

But i have 4 devices with still old bad status, must i reenroll or any db cleanup need or only wait a little bit?

2017-11-07_08h57_53.png

Link to comment
Share on other sites

Guest
This topic is now closed to further replies.
 Share

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...