Centos detects drives wrong deadly slow

[aus-city]

I just installed centos 5.5 x64

It's running raid1 via mdadm

The CPU load is constantly 6 to 7 on quad 3ghz, just seems bad.

After installing and online load jumps to 12 to 15.

I been chasing this seems on disc use it's slow.

I just noticed drives are wrongly detected

They are all sata, but the first two it says are hda1 and hdc1. It detects sda1 but this is my spare drive.

The raid runs on the first two hence the problem.

How can I fix this? It used to run fedora perfectly and detected them all as sata.

What's worse it's set hda1 and hdc1 as non dma.

I read change bios settings? Also I can't kill the install it's now got sites running


Thanks!

[aus-city]

Solved.

I played with the bios its now booting sda through sdc

[mikeshinn]

Might want to leave a tip for what you changed in your BIOS in case anyone else runs into this, its subtle enough I could see someone giving up.

[aus-city]

Yes - it depends on your bios.

Basically against sata in the bios set acpi.

Some bios are bad and don't support advanced settings against sata.

In that case use the kernel options to disable ide and then sata loads.

[mikeshinn]

Heres an interesting question, did you have to do that for Fedora (change the BIOS)? If so, then I wonder if its not so much a kernel issue/driver issue but something else in the tool chain?

[aus-city]

Hi Mike,

No that was why it took me a while to track down.

Fedora mounted all 3 drives as sda to sdc

It's something to do with centos kernel and it miss detecting sata drives as pata.

Sometimes you have to resort to loading kernel specifying no IDE.

Luckily my bios had the advanced settings. What was good did not have to touch grub or anything it fired up properly and the md0 and md1 were still synced.

Machine load back to where it was in fedora.

[mikeshinn]

That is interesting, I wonder if its the driver name? Centos seems to map to older drivers, in fact this can be maddening if you use a newer kernel than what centos uses with really new hardware, it wont "see" it, but Fedora will because of the more up to date driver name mappings. Both centos and fedora can see/use it, but centos/rhel sets up modprobe.conf with the older names which are wrong (so initrd gets created with older names and doesnt work/work properly).

Which kernels have you tried?

[aus-city]

Just the latest out of certos plus.

I would love an asl kernel but need x windows support

[mikeshinn]

You might be able to run X, just enable privileged I/O:

https://www.atomicorp.com/wiki/index.ph ... l.28.29_by

X might want to do some other silly things, so give it a shot.

[premierhosting]

Any chance this stuff relates to my server hangs possibly related to raid1 syncing?

http://atomicorp.com/forums/viewtopic.php?f=1&t=4847

??

[aus-city]

Certainly is possible. What config are your raid drives? Are they hardware or software?

[premierhosting]

Software, how would I report the config?

[aus-city]

Ah same as me

Okay I assume your running a pair of SATA drives?

My issue is centos said they were PATA and their performance was shocking, take a week to sync!

What's the output of fdisk -l

If you see /dev/sda

Your sweet

If you see /dev/hda

You got bios issues. I can help you when I am home you can force the kernel to disable IDE devices and they will get detected as SATA

Good news is your grub and boot is unaffected, machine will run without issues.

[premierhosting]

Code: Select all

[root@server1 ~]# fdisk -l

Disk /dev/sda: 500.1 GB, 500107862016 bytes
255 heads, 63 sectors/track, 60801 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

   Device Boot      Start         End      Blocks   Id  System
/dev/sda1               1         487     3911796   fd  Linux raid autodetect
/dev/sda2             488         731     1959930   82  Linux swap / Solaris
/dev/sda3             732       60801   482512275   fd  Linux raid autodetect

Disk /dev/sdb: 500.1 GB, 500107862016 bytes
255 heads, 63 sectors/track, 60801 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

   Device Boot      Start         End      Blocks   Id  System
/dev/sdb1               1         487     3911796   fd  Linux raid autodetect
/dev/sdb2             488         731     1959930   82  Linux swap / Solaris
/dev/sdb3             732       60801   482512275   fd  Linux raid autodetect

Disk /dev/md1: 4005 MB, 4005560320 bytes
2 heads, 4 sectors/track, 977920 cylinders
Units = cylinders of 8 * 512 = 4096 bytes

Disk /dev/md1 doesn't contain a valid partition table

Disk /dev/md3: 494.0 GB, 494092484608 bytes
2 heads, 4 sectors/track, 120628048 cylinders
Units = cylinders of 8 * 512 = 4096 bytes

Disk /dev/md3 doesn't contain a valid partition table

[aus-city]

Your drives are detected and mounted as SATA.

What's up with your software raid?

Run

cat /proc/mdstat

Post the result please, that will tell what's happening with your raid.