no more MTRRs available

[Michael Schulte zur Sur]

Hi,

we have a redhat 7.2 installation and a bit of a problem, we can't trace. One thing is, that you find above message in the log. Is that serious? The content of /proc/mtrr is down under.

Any comment? thanks,

Michael
reg00: base=0x00000000 ( 0MB), size=2048MB: write-back, count=1
reg01: base=0x400020000 (16384MB), size= 64KB: uncachable, count=1
reg02: base=0x400040000 (16384MB), size= 64KB: uncachable, count=1
reg03: base=0x400060000 (16384MB), size= 64KB: uncachable, count=1
reg04: base=0x400080000 (16384MB), size= 64KB: uncachable, count=1
reg05: base=0x4000a0000 (16384MB), size= 64KB: uncachable, count=1
reg06: base=0x4000c0000 (16384MB), size= 64KB: uncachable, count=1
reg07: base=0x4000e0000 (16384MB), size= 64KB: uncachable, count=1

[Steven E. Protter]

This message has been seen before and appears serious:

Other threads:
http://lists.suse.com/archive/suse-amd64/2003-Dec/0099.html

http://www.uwsg.iu.edu/hypermail/linux/kernel/0301.3/1051.html

http://www.rage3d.com/board/showthread.php?threadid=33736241

http://www.cs.helsinki.fi/linux/linux-kernel/2003-04/0709.html

It appears to be related to BIOS and perhaps memory amounts greater than 4 GB. Tedious as it is you'll have to read about these threads to determine the seriousness of the problem.

My search:

http://www.google.com/search?hl=en&lr=&ie=UTF-8&q=+%22size%3D2048MB%3A+write-back%2C+count%3D1%22&btnG=Search

SEP

[Jeroen Peereboom]

L.S.

I received this messages when starting X (default runlevel is text based).

My /proc/mtrr looks exactly the same as yours.
After reading some internet articles, I felt I could safely ignore these....

But now I feel I have to read SEP's links, and maybe ask HP / RedHat for an opinion.

JP

[Michael Schulte zur Sur]

SEP,

it indeed happened as my collegue tried for the first time to create an oracle database with the graphics tool dbca.
Can I provide any other info to give you an idea, what the problem may be? I have searched with google myself but since I do not have a clue, what to look for, this was too frustrating.

thanks,

Michael

[Steven E. Protter]

This colleague, was he using X windows from a workstation or directly connected to the system via a graphics card?

Based on further analysis and your information, I would speculate that there may be a memory problem or bug that could be fixed by upgrading the kernel. up2date if possible.

I'd also run a check and make sure all the hardware including graphics card and memory is solid.

I really have nothing further, other than the possibility that kernel and other changes required for the oracle install triggered the problem.

SEP

[Michael Schulte zur Sur]

SEP,

he used Reflection for X on his pc. What kind of check are you referring to? As far as I know, there are no changes recently, however noone worked with X.
We are all bloody penguins here! ;-)

thanks,

Michael

[Steven E. Protter]

I have a cd from Dell that lets me boot my boxes and run hardware diagnosis. As crazy as that sounds, its easier for me to take a box offline let the "cluster" failover and work that way.

Of course, there are plenty of itrc posts that say how to check various components, but I'd still consider using a cd from the server maker that lets you check everything.

Disks:
http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=108855

This post is useless but is hilarious, just case you need a laugh.

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=8088

Some useful information in this one:
http://forums1.itrc.hp.com/service/forums/bizsupport/questionanswer.do?threadId=581102

How to check linux for i/o problems
http://search.hp.com/redirect.html?url=http%3A//forums1.itrc.hp.com/service/forums/questionanswer.do%3FthreadId%3D217485&qt=linux+hardware+check&hit=12

More on finding bad disk sectors
http://search.hp.com/redirect.html?url=http%3A//forums1.itrc.hp.com/service/forums/questionanswer.do%3FthreadId%3D493899&qt=linux+hardware+check&hit=10

This is going to have to be enough from me. Production problem.

SEP

[Jeroen Peereboom]

Michael,

as a sidestep, if you try to trace dbca: dbca is VERY sensitive to the slightest disturbance (space, tab, empty line, ...) in Oracle config files (like listener, sqlnet, ...). I only managed to get it running after creating a database manually first...

JP.

[tommy_28]

I also encontered this problem in Proliant DL380g3/360g3/ML570g2 Redhat as3.0U1&U2, mostly with iLO installation, at the phase starting X, it said no more MTRRs avalable, X server terminated and system reboot. This is not good.

the /proc/mtrr info on ml570 are:

[root@ml570-2 root]# cat /proc/mtrr
reg00: base=0xf5d00000 (3933MB), size= 1MB: uncachable, count=1
reg01: base=0xf5e00000 (3934MB), size= 2MB: uncachable, count=1
reg02: base=0xf6000000 (3936MB), size= 32MB: uncachable, count=1
reg03: base=0xf8000000 (3968MB), size= 128MB: uncachable, count=1
reg04: base=0x00004000 ( 0MB), size= 4KB: uncachable, count=1
reg05: base=0x00005000 ( 0MB), size= 4KB: uncachable, count=1
reg06: base=0x00006000 ( 0MB), size= 4KB: uncachable, count=1
reg07: base=0x00007000 ( 0MB), size= 4KB: uncachable, count=1

any method to avoid this?

Thx.