cpu consumption

cpu consumption

[Gilberto Diaz]

Hello Everybody

   I have checked the list archive but I could not find any message that
help. So I'm asking for help here. I have a Sun storege with 12 sata
hard disks connected to a normal PC using a QLogic ISP2300 host adapter
card. I have configured 6 raid1, (2 hard disk each). I excecuted the
following commands to create all of the them

mdadm --create --verbose /dev/md1 --level=raid1 --raid-devices=2
/dev/sda /dev/sdb
mdadm --create --verbose /dev/md2 --level=raid1 --raid-devices=2
/dev/sdc /dev/sdd
mdadm --create --verbose /dev/md3 --level=raid1 --raid-devices=2
/dev/sde /dev/sdf
mdadm --create --verbose /dev/md4 --level=raid1 --raid-devices=2
/dev/sdg /dev/sdh
mdadm --create --verbose /dev/md5 --level=raid1 --raid-devices=2
/dev/sdi /dev/sdj
mdadm --create --verbose /dev/md6 --level=raid1 --raid-devices=2
/dev/sdk /dev/sdl

Here is my /etc/mdadm/mdadm.conf file

DEVICE /dev/sd[abcdefghijkl] /dev/md1 /dev/md2 /dev/md3 /dev/md4
/dev/md5 /dev/md6
ARRAY /dev/md1 devices=/dev/sda,/dev/sdb
ARRAY /dev/md2 devices=/dev/sdc,/dev/sdd
ARRAY /dev/md3 devices=/dev/sde,/dev/sdf
ARRAY /dev/md4 devices=/dev/sdg,/dev/sdh
ARRAY /dev/md5 devices=/dev/sdi,/dev/sdj
ARRAY /dev/md6 devices=/dev/sdk,/dev/sdl

   The problem is that the following proccesses are using a lot of cpu
time.

    md1_raid1
    md1_resync
    ......
    md6_raid1
    md6_resync

    Here is a sample of the uptime command

17:54:16 up  5:48,  2 users,  load average: 5.02, 5.05, 5.26

    Does anybody have an idea what is the problem? Thank a lot in advance

Best regards
Gilberto Diaz

Re: cpu consumption

[Gordon Henderson]

On Mon, 23 Jan 2006, Gilberto Diaz wrote:

>    The problem is that the following proccesses are using a lot of cpu
> time.
>
>     md1_raid1
>     md1_resync
>     ......
>     md6_raid1
>     md6_resync
>
>     Here is a sample of the uptime command
>
> 17:54:16 up  5:48,  2 users,  load average: 5.02, 5.05, 5.26
>
>     Does anybody have an idea what is the problem? Thank a lot in advance

to me, it looks like the "problem" is that the RAID system is still
creating the array and syncing all the partitions up. Once it's finished
your system will be idle.

I bet it's stopped doing it by the time you get this email... What does
the output of

  cat /proc/mdstat

look like?

Gordon

Re: cpu consumption

[Gilberto Diaz]

Thanks Gordon, Guy, Michel and Neil, your advises were very useful.

    Gordon was right, all raids are still syncing (Bellow part of the
/proc/mdstat file).  Guy you were right too, the command "uptime" does
not display the cpu load directly, sorry about that.


Best regards
Gilberto

----------------------------------------------------------------------------------------------------------------------------------------------------
Personalities : [linear] [raid0] [raid1] [raid5] [multipath] [raid6]
[raid10] [faulty]
 
md6 : active raid1 sdk[0] sdl[1]
      244190144 blocks [2/2] [UU]
      [================>....]  resync = 83.4% (203849152/244190144)
finish=5.0min spee
d=131933K/sec
     
md5 : active raid1 sdi[0] sdj[1]
      244190144 blocks [2/2] [UU]
      [================>....]  resync = 83.8% (204684480/244190144)
finish=5.1min spee
d=127124K/sec
     
.
.
.

--------------------------------------------------------------------------------------------------------------------------------------------------------------


Gordon Henderson wrote:

>On Mon, 23 Jan 2006, Gilberto Diaz wrote:
>
>  
>
>>   The problem is that the following proccesses are using a lot of cpu
>>time.
>>
>>    md1_raid1
>>    md1_resync
>>    ......
>>    md6_raid1
>>    md6_resync
>>
>>    Here is a sample of the uptime command
>>
>>17:54:16 up  5:48,  2 users,  load average: 5.02, 5.05, 5.26
>>
>>    Does anybody have an idea what is the problem? Thank a lot in advance
>>    
>>
>
>to me, it looks like the "problem" is that the RAID system is still
>creating the array and syncing all the partitions up. Once it's finished
>your system will be idle.
>
>I bet it's stopped doing it by the time you get this email... What does
>the output of
>
>  cat /proc/mdstat
>
>look like?
>
>Gordon
>
>  
>

RE: cpu consumption

[Guy]

} -----Original Message-----
} From: linux-raid-owner@vger.kernel.org [mailto:linux-raid-
} owner@vger.kernel.org] On Behalf Of Gilberto Diaz
} Sent: Monday, January 23, 2006 5:28 PM
} To: linux-raid@vger.kernel.org
} Subject: cpu consumption
} 
} Hello Everybody
} 
}    I have checked the list archive but I could not find any message that
} help. So I'm asking for help here. I have a Sun storege with 12 sata
} hard disks connected to a normal PC using a QLogic ISP2300 host adapter
} card. I have configured 6 raid1, (2 hard disk each). I excecuted the
} following commands to create all of the them
} 
} mdadm --create --verbose /dev/md1 --level=raid1 --raid-devices=2
} /dev/sda /dev/sdb
} mdadm --create --verbose /dev/md2 --level=raid1 --raid-devices=2
} /dev/sdc /dev/sdd
} mdadm --create --verbose /dev/md3 --level=raid1 --raid-devices=2
} /dev/sde /dev/sdf
} mdadm --create --verbose /dev/md4 --level=raid1 --raid-devices=2
} /dev/sdg /dev/sdh
} mdadm --create --verbose /dev/md5 --level=raid1 --raid-devices=2
} /dev/sdi /dev/sdj
} mdadm --create --verbose /dev/md6 --level=raid1 --raid-devices=2
} /dev/sdk /dev/sdl
} 
} Here is my /etc/mdadm/mdadm.conf file
} 
} DEVICE /dev/sd[abcdefghijkl] /dev/md1 /dev/md2 /dev/md3 /dev/md4
} /dev/md5 /dev/md6
} ARRAY /dev/md1 devices=/dev/sda,/dev/sdb
} ARRAY /dev/md2 devices=/dev/sdc,/dev/sdd
} ARRAY /dev/md3 devices=/dev/sde,/dev/sdf
} ARRAY /dev/md4 devices=/dev/sdg,/dev/sdh
} ARRAY /dev/md5 devices=/dev/sdi,/dev/sdj
} ARRAY /dev/md6 devices=/dev/sdk,/dev/sdl
} 
}    The problem is that the following proccesses are using a lot of cpu
} time.
} 
}     md1_raid1
}     md1_resync
}     ......
}     md6_raid1
}     md6_resync
} 
}     Here is a sample of the uptime command
} 
} 17:54:16 up  5:48,  2 users,  load average: 5.02, 5.05, 5.26

What is the CPU load?  The above does not reflect CPU usage.
I beleve it reports on average outsanding IOs.  With 6 arrays syncing, I
would have expected 6 or 12.

Run "top", or "sar -u 10 10" to see CPU usage.

Guy
} 
}     Does anybody have an idea what is the problem? Thank a lot in advance
} 
} Best regards
} Gilberto Diaz
} -
} To unsubscribe from this list: send the line "unsubscribe linux-raid" in
} the body of a message to majordomo@vger.kernel.org
} More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: cpu consumption

[Neil Brown]

On Monday January 23, gilberto@ula.ve wrote:
> 
> Here is my /etc/mdadm/mdadm.conf file
> 
> DEVICE /dev/sd[abcdefghijkl] /dev/md1 /dev/md2 /dev/md3 /dev/md4
> /dev/md5 /dev/md6
> ARRAY /dev/md1 devices=/dev/sda,/dev/sdb
> ARRAY /dev/md2 devices=/dev/sdc,/dev/sdd
> ARRAY /dev/md3 devices=/dev/sde,/dev/sdf
> ARRAY /dev/md4 devices=/dev/sdg,/dev/sdh
> ARRAY /dev/md5 devices=/dev/sdi,/dev/sdj
> ARRAY /dev/md6 devices=/dev/sdk,/dev/sdl

This is a poor choice for an mdadm.conf file.
The device names of 'SCSI' devices are prone to change with changing
system configuration.  If something disappears or is added, all the
names could change, and this will cease to work.
I would strongly recommend putting UUIDs in there.

  echo 'DEVICE /dev/sd[a-l]' > /etc/mdadm/mdadm.conf
  mdadm -Ds | grep -v devices >> /etc/mdadm/mdadm.conf

NeilBrown

Re: cpu consumption

[Mitchell Laks]

On Monday 23 January 2006 10:05 pm, Neil Brown wrote:

>
> This is a poor choice for an mdadm.conf file.
> The device names of 'SCSI' devices are prone to change with changing
> system configuration.  If something disappears or is added, all the
> names could change, and this will cease to work.
> I would strongly recommend putting UUIDs in there.
>
>   echo 'DEVICE /dev/sd[a-l]' > /etc/mdadm/mdadm.conf
>   mdadm -Ds | grep -v devices >> /etc/mdadm/mdadm.conf
This last line gives nothing because  -v switch selects lines not having 
devices, but the other line has num-devices which also matches:
ie
C1:~# mdadm -Ds
ARRAY /dev/md3 level=raid1 num-devices=2 
UUID=209ab7c8:d37fee49:3a63af71:99193e75
   devices=/dev/sde1
ARRAY /dev/md2 level=raid1 num-devices=2 
UUID=03dc44fa:606063e2:83e69037:50d00eb5
   devices=/dev/sdd1
ARRAY /dev/md1 level=raid1 num-devices=2 
UUID=88433668:ffa6265d:5ed3569b:9ef1d0f8
   devices=/dev/sdc1
ARRAY /dev/md0 level=raid1 num-devices=2 
UUID=5f001583:253365c1:1092ccdf:b052e938
   devices=/dev/sdb1,/dev/sda1

thus 
C1:~# mdadm -Ds|grep -v devices
C1:~#   

while you must want
C1:~# mdadm -Ds |grep  num-devices
ARRAY /dev/md3 level=raid1 num-devices=2 
UUID=209ab7c8:d37fee49:3a63af71:99193e75
ARRAY /dev/md2 level=raid1 num-devices=2 
UUID=03dc44fa:606063e2:83e69037:50d00eb5
ARRAY /dev/md1 level=raid1 num-devices=2 
UUID=88433668:ffa6265d:5ed3569b:9ef1d0f8
ARRAY /dev/md0 level=raid1 num-devices=2 
UUID=5f001583:253365c1:1092ccdf:b052e938

(excuse the  line wrap by kmail)

thanks
Mitchell Laks




>
> NeilBrown
> -
> To unsubscribe from this list: send the line "unsubscribe linux-raid" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

cpu consumption

[Raz Ben-Jehuda(caro)]

Hello Jeff.
Here is my problem.

When I issue a READ operation over a SATA disk in linux i have
huge cpu consumption , about 90% cpu usage.

I am using MAXTOR disks in a supermicro machine.
the cpu test was made when i "raw read" from the DISK using DIRECT_IO.

XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

this is my dmesg.
ata1: SATA max UDMA/133 cmd 0xE900 ctl 0xEA02 bmdma 0xED00 irq 10
ata2: SATA max UDMA/133 cmd 0xEB00 ctl 0xEC02 bmdma 0xED08 irq 10
ata1: dev 0 cfg 49:2f00 82:7c6b 83:7f09 84:4673 85:7c69 86:3e01 87:4663 88:207f
ata1: dev 0 ATA, max UDMA/133, 490234752 sectors: lba48
ata1: dev 0 configured for UDMA/133
scsi0 : ata_piix
ata2: dev 0 cfg 49:2f00 82:7c6b 83:7f09 84:4673 85:7c69 86:3e01 87:4663 88:207f
ata2: dev 0 ATA, max UDMA/133, 490234752 sectors: lba48
ata2: dev 0 configured for UDMA/133
scsi1 : ata_piix
  Vendor: ATA       Model: Maxtor 7L250S0    Rev: BANC
  Type:   Direct-Access                      ANSI SCSI revision: 05
  Vendor: ATA       Model: Maxtor 7L250S0    Rev: BANC
  Type:   Direct-Access                      ANSI SCSI revision: 05
SCSI device sda: 490234752 512-byte hdwr sectors (251000 MB)
SCSI device sda: drive cache: write back
SCSI device sda: 490234752 512-byte hdwr sectors (251000 MB)
SCSI device sda: drive cache: write back
 /dev/scsi/host0/bus0/target0/lun0: p1 p2
Attached scsi disk sda at scsi0, channel 0, id 0, lun 0
SCSI device sdb: 490234752 512-byte hdwr sectors (251000 MB)
SCSI device sdb: drive cache: write back
SCSI device sdb: 490234752 512-byte hdwr sectors (251000 MB)
SCSI device sdb: drive cache: write back
 /dev/scsi/host1/bus0/target0/lun0:
Attached scsi disk sdb at scsi1, channel 0, id 0, lun 0
Attached scsi generic sg0 at scsi0, channel 0, id 0, lun 0,  type 0
Attached scsi generic sg1 at scsi1, channel 0, id 0, lun 0,  type 0
mice: PS/2 mouse device common for all mice
i2c /dev entries driver
md: linear personality registered as nr 1
md: raid0 personality registered as nr 2
md: raid1 personality registered as nr 3
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX


Thank you.
--
Raz

Re: cpu consumption

[Jeff Garzik]

Raz Ben-Jehuda(caro) wrote:
> Hello Jeff.
> Here is my problem.
> 
> When I issue a READ operation over a SATA disk in linux i have
> huge cpu consumption , about 90% cpu usage.
> 
> I am using MAXTOR disks in a supermicro machine.
> the cpu test was made when i "raw read" from the DISK using DIRECT_IO.

I dunno... do some profiling (in-kernel profiler, or oprofile) and see 
where the CPU is being consumed.

Are you also seeing a lot of interrupts?

	Jeff

Re: cpu consumption

[Tejun Heo]

Jeff Garzik wrote:
> Raz Ben-Jehuda(caro) wrote:
> 
>> Hello Jeff.
>> Here is my problem.
>>
>> When I issue a READ operation over a SATA disk in linux i have
>> huge cpu consumption , about 90% cpu usage.
>>
>> I am using MAXTOR disks in a supermicro machine.
>> the cpu test was made when i "raw read" from the DISK using DIRECT_IO.
> 
> 
> I dunno... do some profiling (in-kernel profiler, or oprofile) and see 
> where the CPU is being consumed.
> 
> Are you also seeing a lot of interrupts?
> 
>     Jeff
> 
> 

Jeff, this one is resolved.  It turned out that Raz mistook wait time to 
be busy.

Hmmm... marc doesn't show any mesage after 2005-11-25.  This is is 
resolved in a thread on linux-ide named "CPU usage is very high while 
io'ing a sata disk" from Raz.  The first message of the thread is posted 
on 2005-11-30 06:23 +0900.

-- 
tejun