Just reread your post again.... Actually, "cat /proc/partitions", if the said partitions/disks are still being seen by kernel, you will need some sort of system scan to get rid of them, or a reboot. I have a SAN, and qlogic card as HBA, qlogic has a tool for scanning for non-existent partitions and remove them... http://filedownloads.qlogic.com/files/ms/56615/readme_dynamic_lun_22.html and with qlogic, I could have removed the non-existent LUNs via the option refresh: -r, --refresh                         To refresh, that is remove LUNs that are lost                         use the options "-r|--refresh". This will                         remove the LUNs which no more exist. How were these disks attached to your system?

----- Original Message ----- From: Philippe Naudin philippe.naudin@supagro.inra.fr To: CentOS mailing list centos@centos.org Cc: Sent: Friday, October 28, 2011 9:26 AM Subject: Re: [CentOS] read failed after messages of non existing harddisks

Le ven 28 oct 2011 14:08:50 CEST, Götz Reinicke a écrit:

Hi,

some time ago I removed some physical disks from a server and now I'm still getting dmesg messages like:

sd 0:2:2:0: SCSI error: return code = 0x00040001 end_request: I/O error, dev sdc, sector 0

And all lvm tools still grumbel about that disks too:

/dev/sdb: read failed after 0 of 4096 at 0: Eingabe-/Ausgabefehler   /dev/sdb: read failed after 0 of 4096 at 1746969493504: Eingabe-/Ausgabefehler

How may I tell the lvm and the system, that it is ok that this disk do not exist any more?

If a software RAID has been removed, don't forget :   mdadm --misc --zero-superblock /dev/sdX where sdX is every disk that was part of the RAID.

I would assume you have rebooted since removing the drives, in which case they should not be seen by "cat /proc/partition", which is a representation of scanned disks during bootup, unless you have other drives still connected that you are not aware of? "cat /proc/scsi/scsi" should help in narrowing down which devices you are actually looking at.

To bridge the gap when what partitions and scsi/scsi is showing, install package (if not there already) "lsscsi" and run command lsscsi. Sample output: $ lsscsi [0:0:6:0]    process PE/PV    1x3 SCSI BP      1.1   -       [0:2:0:0]    disk    MegaRAID LD 0 RAID1  139G 422D  /dev/sda [1:0:0:0]    mediumx COMPAQ   MSL5000 Series   0520  -       [1:0:1:0]    tape    COMPAQ   SDLT320          5F5F  /dev/st0 [1:0:2:0]    tape    COMPAQ   SDLT320          5F5F  /dev/st1

----- Original Message ----- From: Götz Reinicke goetz.reinicke@filmakademie.de To: CentOS mailing list centos@centos.org Cc: Sent: Monday, October 31, 2011 5:37 AM Subject: Re: [CentOS] read failed after messages of non existing harddisks

Hi,

a cat /proc/partitions shows the two disks. They were direct connected to an scsi port. (Old ide 2 scsi Raids).

I disconnected tham a long time ago and never worried about the errors, as the system was up and running very smooth.

But now I connect other iscsi systems and the dmes and log is not that easy to monitor with all the unneded messages ...

Anny idea, how to remove that partitions with the redhat tools?

Regards . Götz

Am 28.10.11 18:10, schrieb Richard Mollel:

Just reread your post again.... Actually, "cat /proc/partitions", if the said partitions/disks are still being seen by kernel, you will need some sort of system scan to get rid of them, or a reboot. I have a SAN, and qlogic card as HBA, qlogic has a tool for scanning for non-existent partitions and remove them... http://filedownloads.qlogic.com/files/ms/56615/readme_dynamic_lun_22.html and with qlogic, I could have removed the non-existent LUNs via the option refresh: -r, --refresh                         To refresh, that is remove LUNs that are lost                         use the options "-r|--refresh". This will                         remove the LUNs which no more exist. How were these disks attached to your system?

----- Original Message ----- From: Philippe Naudin philippe.naudin@supagro.inra.fr To: CentOS mailing list centos@centos.org Cc: Sent: Friday, October 28, 2011 9:26 AM Subject: Re: [CentOS] read failed after messages of non existing harddisks

Le ven 28 oct 2011 14:08:50 CEST, Götz Reinicke a écrit:

...

Hi,

some time ago I removed some physical disks from a server and now I'm still getting dmesg messages like:

sd 0:2:2:0: SCSI error: return code = 0x00040001 end_request: I/O error, dev sdc, sector 0

And all lvm tools still grumbel about that disks too:

/dev/sdb: read failed after 0 of 4096 at 0: Eingabe-/Ausgabefehler     /dev/sdb: read failed after 0 of 4096 at 1746969493504: Eingabe-/Ausgabefehler

How may I tell the lvm and the system, that it is ok that this disk do not exist any more?

If a software RAID has been removed, don't forget :   mdadm --misc --zero-superblock /dev/sdX where sdX is every disk that was part of the RAID.