This content has been marked as final. Show 5 replies
Is it seen by linux as a hard drive (e.g. /dev/sdb1) or is it a logical volume (e.g. /dev/mapper/VolGroup_ID_18405-LogVolU02) ?
I think the easier way is partitioning this free space and accessing in "raw mode" to the partition
If you disk is /dev/sda , you will have your 50Gb of linux in /dev/sda1 and a new unformatted partition on /dev/sda2
You must grant permissions to oracle by "chown oracle:oinstall /dev/sda2" and you can access to /dev/sda2 in raw mode.
Edited by: clementeOTN on 01-feb-2013 13:55
I suggest you to partition the free space, because i think is the easiest and safest way to access to this disk without destroy the linux data.
If it wasn't a hard drive (for example a big storage lun) you would access directly to the device, but at the end, is the same to acces to a entire disk /dev/sdX than acces to a /dev/sdX1 partition because you will make an raw access to an device whatever you call it
Don't do it. ASM requires more than one device. and making multiple "devices" out of one physical device by partitioning may be okay for testing, but I wouldn't use it like that for production. You also need to make sure all devices in an ASM diskgroup are EXACTLY the same size. this is an imperative. Also, if you do use it for testing, do not expect any real performance out of that configuration.
You also need to make sure all devices in an ASM diskgroup are EXACTLY the same size. this is an imperative.Please correct me if I'm wrong, but isn't Oracle ASM capacity based and writes to a disks in a round-robin fashion? So if you have disks of a different size, I/O and usage will stop to be balanced as some of the disks run out of free space. To use the same size is therefore recommended, but not imperative.
To the OP: Raw device support under RHEL 5 and derivatives is depreciated. ASM data redundancy is typically achieved by mirroring your data among separate physical drives, which was already mentioned.