RHEL5 boot issue - raid not found after updates

[Anonymous]

I updated RHEL5 after recent security and bug alerts and now can't boot the server. It does not recognize/find the raid group. I know there is a specific driver for our controller, however it does not seem to work on the latest updates.

HP ProLiant DL320 G6 Server series
HP Smart Array B110i SATA RAID Controller
RHEL V5 -
PREVIOUS level - 2.6.18-194.17.1.el5xen
UPDATED level - 2.6.18-194.26.1.el5xen

I am stuck at this point as I can't get the server to boot even by selecting the previous level.

FYI - not very skilled at Linux - sorry.

Thanks, Jeff

[Chhaya_Z]

Hi Jeff,

Have you checked the status of the RAID in the BIOS Advance settings?

During boot where does it fail? what is the last message on the console?

[Benoy Daniel]

Jeff,

Please go to ILO and see the log if there is anything unusual. Also check whether the RAID controller is detected and disk is visible or not.

[Matti_Kurkela]

HP Smart Array B110i is a special case: it is not a "true hardware RAID" like most other SmartArray models.

The standard kernel already has a driver which can handle the AHCI SATA controllers, which the SmartArray B110i is based on. But the standard driver does not understand the SmartArray RAID configuration - only the proprietary SmartArray B110i driver can do that.

So the B110i driver needs to *override* the standard RHEL5 AHCI driver. If this is not done correctly, the RAID array will not work, and the standard AHCI driver will claim the hardware instead.

According to the Release Notes of the hpahcisr driver, the latest supported RHEL5 kernel version is 2.6.18-194.8.1.el5.

You may have installed the 2.6.18-194.17.1.el5xen kernel without rebooting the system, and thus remained unaware of this fact. If your bootloader menu has multiple older kernel versions available, one of those might allow you to boot. (It should have, unless someone has taken the trouble to remove the old versions.)

But if those kernel versions you mentioned are all that's currently available on the system, you'll need a more involved rescue procedure:

Download the Driver Diskette package from here:

(64-bit version:)
http://h20000.www2.hp.com/bizsupport/TechSupport/SoftwareDescription.jsp?lang=en&cc=us&prodTypeId=329290&prodSeriesId=3958193&prodNameId=3958194&swEnvOID=4004&swLang=13&mode=2&taskId=135&swItem=MTX-9e896dcbc5744a4cb456bdfff8

(32-bit version:)
http://h20000.www2.hp.com/bizsupport/TechSupport/SoftwareDescription.jsp?lang=en&cc=us&prodTypeId=329290&prodSeriesId=3958193&prodNameId=3958194&swEnvOID=4006&swLang=13&mode=2&taskId=135&swItem=MTX-67c4d29336be4033acb6e636b6

Despite the name, the Driver Diskette package can also be used with an USB stick or other removable media: just save the hpahcisr-.rhel5..dd file to the USB stick as-is.

Also download the latest driver RPM from HP and the appropriate kernel version RPM from RedHat Network web interface.

Boot the system into rescue mode from the RHEL5 installation media, using the driver diskette in the boot process (i.e. type "linux rescue dd" in the bootloader prompt).
When prompted, insert the media that contains the driver diskette package.

This should allow you to access the system in command-line mode, with the installed root filesystem mounted to /mnt/sysimage.

Run "chroot /mnt/sysimage /bin/bash": it will allow you to operate the system as if it were booted to single-user mode normally, with all the filesystems in their usual locations. Download and install a supported kernel version and the hpahcisr driver RPM. If the installation procedure is succesful, your system should again be able to boot from the RAID.

After you've fixed this problem, do whatever is appropriate at your site to make this extremely visible to sysadmins: "DO NOT UPGRADE THE KERNEL OF THIS SYSTEM UNLESS YOU ALSO UPGRADE THE HPAHCISR RPM TO SUPPORT THE UPGRADED KERNEL VERSION AT THE SAME TIME".

MK