Hello group,I have a problem with an ASM disk group on a freshly installed server. We are using RedHat 5.3 (64bit) and Oracle 11.1.0.7 for clusterware and ASM. Initially the disk group creation went fine, but after a couple of days we encountered metadata corruption which manifests itself in the following:
ORA-15196: invalid ASM block header [kfc.c:9133] [endian_kfbh] [3] [2147483648] [17 != 1]
We based the disk group on raw devices, using udev to map the logical devices created by HDLM (sddlmaxx if memory serves me right).
The storage admins have run /sbin/dlmcfgmgr -r after adding in more luns to prompt the generation of the logical device names - could that have had an impact on the existing ones?
Does anyone here have experience with HDLM 6.x and Linux? Every hit I get on the WWW seems to be related to AIX. We might consider moving to ASMLib, but I don't want to introduce another layer of software to the stack if I don't have to.
NB: there are 2 bugs related to HDLM but they seem non-public: Bug 4963316 - ASM CANNOT DISCOVER PHYSICAL DISKS CREATED BY HDML; Bug 8273208 - PLEASE ADD MKNOD AS A OPTION WHEN USING HDLM AND ASM Thanks for any tips, Martin -- //www.freelists.org/webpage/oracle-l