Topic: du continually using CPU

[hilljockey]

Hello,

I have SME server with kernel 2.6.9-78.0.22.EL.

This kernel was installed during a recent upgrade.

Process du has been going continually since the first tape backup (LTO Ultirum).  Have rebooted and it is still going.  I keep getting alerts like "Your processor is 98.22 % loaded".

Top shows this to be a root process, the wa (waiting) % is usually high.  The process number keeps increasing, so it must be a new instance every time.

Has anyone has a similar problem?  Is there more information I could/should provide to contribs?  Is it a bug?

I'm a bit light on with nix experience so pointers on how to extract information about the problem would be handy too... (sorry).

[cactus]

I keep getting alerts like "Your processor is 98.22 % loaded".

Where do you observe these alerts? AFAIK they are not part of a vanilla SME Server installation.

Top shows this to be a root process, the wa (waiting) % is usually high.  The process number keeps increasing, so it must be a new instance every time.

Not sure if the assumption on the new instance is true...

Has anyone has a similar problem?  Is there more information I could/should provide to contribs?  Is it a bug?

I am nut sure on that, but problems and issues should be tracked in the bugtracker as is stated with every post you post to the forums.

I'm a bit light on with nix experience so pointers on how to extract information about the problem would be handy too... (sorry).

AFAIK high processor loads are not abnormal for Linux based systems, Linux tends to use the resources when it needs them, but with a far better priority system, preventing your machine from lock-up (like for instance windows machines).

[hilljockey]

Thanks Cactus.  I will put it on the bug tracker.  First time I've posted here and right as I was clicking the post button I saw the bit about bug report being better (in red no less!).

SME admin is sending me the alerts via email.

Our server is a couple of years old and this sort of high CPU use being reported is usually only during a tape backup concurrent with ClamAV.

This recent issue of continual high CPU is only since changing to the newer kernel.  The server is actually still doing it's thing as always, but the HD LED on the front is on constantly.  That makes me a little worried about the hard drives.

Incidentally, the kernel was part of an upgrade, but then we went back a version as our Windows terminal server wouldn't work with VPN from an external windows PC (but OK with TSclient, and, anything would access desktops on the network).  The problem is since moving forward to the most recent kernel so VPN would work.  We had this kernel without the du issue for about a week.

[christian]

It sounds to me that you are using Sme7admin which is a contrib.

It sends those types of alerts and also has an option to use "du -s".

In the server manager under "Administration" you likely have an entry calls "Sme7admin". Click on this and then click the link "Configuration of sme7admin".

About 2/3 of the way done under a heading called "Configuration of hardware sensors" there is an option "Activate du -s". Set it to "off" and then click the save button at the bottom of the page.

[hilljockey]

Thanks christian, problem solved. 

 I will put it on the contribs bugtracker.  We've had that contrib for over 2 years and never seen that problem, the du -s was always enabled (I think).