One of my main machines was runing just fine for the past 8 - 9 months. I've got 6 hard drives on it - 2 - 1 TB drives with 3 raid-1 (/devsda & /dev/sdd) partitions for /boot; vg_opsys; vg_sysswap and the remainder added to the raid-10 defined in vg_sysdata. The remaining 4 hard drives (2 - 1 TB and 2 - 2 TB) used in the construction of the vg_sysdata. One of the 1 TB drives in the RAID-10 device totally failed. I don't have a extra 1 TB disk available, but I do have a 3 TB disk. What I want to do is define a partition the looks like a 1 TB disk (start at 2048 and end at 1710032895) add this to the RAID-10 as /dev/sde1. Then I want to divide the remainder of the 3 TB disk in half and add both to the RAID-10 increasing it's capacity. BTW, all of this is using software raid on CentOS 7.9. Does anyone see potential issues or should I just wait until the 2 - 1 TB disks get here in the mail??? Gene
I would say it is a bit iffy to have a phycical disk with partitions belonging to two distict RAID packs or volume groups. I try to avoid unorthodox setups just in case. Things like you are doing may work but I do not have time to reseach the issues enough to be comfortable and trust a setup that is not what the documentation advices. Performance is sure to suffer, at least. So I would add that 3 TB disk to the failed RAID pack and just leave the rest of it unused. Then when the replacement disk with suitable size arrives, put that in place of the 3 TB disk. But since you already have a setup where disk partitions are set up that way, you could continue the same route and do what you proposed. I have no idea of the possible catastrophes that may or may not cause, however.
two partitions on the same physical disk, spread across the same raid array? firstly, it's going to absolutely thrash the read/write heads. they're going to be all over the place, increases the chances of a disk failure. secondly, if the drive fails, that's two partitions in the raid array gone. depending on exactly which partitions they are in a raid-10, you may just get away with it, much more likely is you can't recover the array at all. i'd follow @Taleman's instructions, just use the 3tb disk as is, just get the raid array rebuild back up to 100%, when you get the new drives, fail the 3tb drive, remove it and replace it with the new drive and let it rebuild again. if you want to increase capacity, just use a 1tb partition of the 3tb disk, and let it rebuild, once done, replace the other drive with a 3tb drive, let it rebuild, then expand the partitions to use the rest of the drives
