Skip to contentSkip to conversation

converting from external log to internal log ... is this possible?

DISCUSSION Joe Landman 3 participants 3 messages Sep 8, 2014

lore
Share & tools
lore
#0 lore raw

Figured I'd ask, looked on the FAQ and didn't see anything there relevant to this. There is a hint of this in xfs_growfs, specifically the -i option, though it is noted that "[NOTE: This option is not implemented]".

--
Joe Landman

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs

#1 lore raw
Quoted 4 lines · Joe Landman

Figured I'd ask, looked on the FAQ and didn't see anything there relevant to this. There is a hint of this in xfs_growfs, specifically the -i option, though it is noted that "[NOTE: This option is not implemented]".

Yep, it's never been implemented. You could do it manually with some very careful xfs_db surgery, if you had enough contiguous freespace available, but it's not at all supported.

Basically, no, you can't do it without getting very far under the hood and directly editing the disk.

-Eric

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs

#2 lore raw
Quoted 6 lines · Eric Sandeen
Quoted 4 lines · Joe Landman

Figured I'd ask, looked on the FAQ and didn't see anything there relevant to this. There is a hint of this in xfs_growfs, specifically the -i option, though it is noted that "[NOTE: This option is not implemented]".

Quoted 6 lines

Yep, it's never been implemented. You could do it manually with some very careful xfs_db surgery, if you had enough contiguous freespace available, but it's not at all supported.

Basically, no, you can't do it without getting very far under the hood and directly editing the disk.

There's nothing amazingly difficult about implementing external-to-internal journal. First allocate the contiguous extent for the new log, then zero it, then freeze the filesystem to bring the external log down to a clean state, atomically modify the superblock on disk to point at the new log, then re-initialise all the log state to point at the new log, update the head and tail to new cycles at the start of the new log, and unfreeze....

Cheers,

Dave.

Signature · 7 lines
Dave Chinner
[email protected]

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs