Kernal 2.6.15-26 Server Memory Leaks and Disk I/O errors

Bug #55692 reported by Moncrieff,Jeffrey
6
Affects Status Importance Assigned to Milestone
linux-source-2.6.15 (Ubuntu)
Invalid
Undecided
Unassigned

Bug Description

we have a server that has (Dapper server LTS ) Kernel 2.6.15-26-server. We have been geting some errors in dmesg and the memory (RAM) has not been allocated correctly . It has 1 Gig or ram and no strees I see Memory been dumped in cache like 931 going n cache and 131 free that is not right
We also have what looks like cve-1859 and also when we put stress on the disk we get Status 0X51 errors. Please see bug#53798

Revision history for this message
Ben Collins (ben-collins) wrote :

Please attach dmesg.

Revision history for this message
Ben Collins (ben-collins) wrote :

Please note that CVE 1859 is already patched in our latest dapper kernel.

Also, describe more completely the behavior you are seeing and the behavior that you are expecting. The numbers you show above meen very little without some context to explain why they are incorrect.

Revision history for this message
Moncrieff,Jeffrey (jeffrey-moncrieff) wrote :

dmesg

3724395.730000] ata1: status=0x51 { DriveReady SeekComplete Error }
[43724436.990000] ata1: no sense translation for status: 0x51
[43724436.990000] ata1: translated ATA stat/err 0x51/00 to SCSI SK/ASC/ASCQ 0x3/11/04
[43724436.990000] ata1: status=0x51 { DriveReady SeekComplete Error }

Then the primary drive shut down
 Free

Revision history for this message
Ben Collins (ben-collins) wrote :

Without some more info on the memory issue, I'll assume there's no security problem.

As for the IO errors, please attach full dmesg output, and describe exactly what steps you take to reproduce the problem.

Revision history for this message
Moncrieff,Jeffrey (jeffrey-moncrieff) wrote :

Today I checked the dmesg. And I notice a error like this

ata1: command 0x25 timeout, stat 0x50 host_stat 0x61 Free looks like this

total used free shared buffers cached
Mem: 1003 532 470 0 157 116
-/+ buffers/cache: 258 744
Swap: 2917 0 2917

I not use a lot of services just apache mysql bind and a postfix MTA. The MTA is not live at moment.
The webserver is not live (preproduction) How come I only have 532 free where I should have about 800.

Revision history for this message
Moncrieff,Jeffrey (jeffrey-moncrieff) wrote :
Revision history for this message
Ben Collins (ben-collins) wrote :

Output of "ps axu > ps.txt" would be helpful.

Revision history for this message
Moncrieff,Jeffrey (jeffrey-moncrieff) wrote :
Revision history for this message
Ben Collins (ben-collins) wrote :

Ok, from what it looks like, most of your memory is used up in cache. Are you doing anything disk intensive?

Please provide output of "lsmod > lsmod.txt"

Revision history for this message
Moncrieff,Jeffrey (jeffrey-moncrieff) wrote :
Revision history for this message
Ben Collins (ben-collins) wrote :

Sorry, need these too:

cat /proc/meminfo > meminfo.txt
cat /proc/slabinfo > slabinfo.txt

Revision history for this message
Moncrieff,Jeffrey (jeffrey-moncrieff) wrote :
Revision history for this message
Moncrieff,Jeffrey (jeffrey-moncrieff) wrote :
Revision history for this message
Ben Collins (ben-collins) wrote :

I don't see anything out of the ordinary. You have a lot of cache used, but again, that's normal, especially if you are doing a lot of disk related things (like web, sql, etc).

The cache isn't "used memory" per se, but it does show up in the free list as being used. If userspace programs need this memory, the kernel will automatically free from the cache.

I think the memory aspect of this bug report is a non-issue. If you want more assistance with the I/O error, please provide a full dmesg with this problem occuring in it.

Revision history for this message
Ben Collins (ben-collins) wrote :

Checked on this, rejecting this bug, since the ATA errors are in another bug report (55692).

Changed in linux-source-2.6.15:
status: Needs Info → Rejected
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers