md bitmaps on 2.6.16.y
I have been testing an md RAID 1 created using mdadm 2.6.1 on a 2.6.16.35 kernel, and the initial results are encouraging. (The 2.6.16.y kernel is a long-term stable branch, unlike most stable kernels, which are maintained only until the next stable branch becomes active.)
I see bitmap-related patches with fixes for later kernels being discussed in the linux-raid mailing list, but I do not see bitmap-related patches in the ChangeLogs for the 2.6.16.y kernel. My hope is that these fixes are only part of the natural churn in the newer kernel branches and are not needed for the 2.6.16.y kernels.
Are there any known issues that would make the use of md bitmaps in a 2.6.16.y kernel problematic? I do see an important RAID 1 fix in 2.6.16.40, so I know that 2.6.16.35 isn't good enough.
Signature · 1 line
Ed L Cashin <[email protected]>
Quoted 12 lines · On Monday March 26, [email protected] wrote:
I have been testing an md RAID 1 created using mdadm 2.6.1 on a 2.6.16.35 kernel, and the initial results are encouraging. (The 2.6.16.y kernel is a long-term stable branch, unlike most stable kernels, which are maintained only until the next stable branch becomes active.)
I see bitmap-related patches with fixes for later kernels being discussed in the linux-raid mailing list, but I do not see bitmap-related patches in the ChangeLogs for the 2.6.16.y kernel. My hope is that these fixes are only part of the natural churn in the newer kernel branches and are not needed for the 2.6.16.y kernels.
Quoted 3 lines
Are there any known issues that would make the use of md bitmaps in a 2.6.16.y kernel problematic? I do see an important RAID 1 fix in 2.6.16.40, so I know that 2.6.16.35 isn't good enough.
While I try to remember to back-port any significant and relevant bugfixes to 2.6.16, I do forget sometimes.. and 2.6.16 isn't really my focus of development.
If you would like to generate a list of bitmap bugfixes in later kernels, I could check and let you know how relevant they are for 2.6.16.
NeilBrown