On Thu, 24 Feb 2011, Henry Chang wrote: > I want to confirm if the following problem is related to Bug #630. > > Step 1. On client A, create a file of 10G by dd. > "ceph -w" shows the total data size increased by 10G. > > Step 2. On client B, touch the file. > > Step 3. On client A, remove the file. > "ceph -w" shows the total data size is NOT decreased. > > Step 4. Umount ceph on client B. > "ceph -w" shows the total data size decreased by 10G.
It is. The bug is that the data doesn't get removed until the inode drops out of the inode cche on client B, and there is currently no hint from the MDS that it has been unlinked, so that takes however long it takes for the VFS to get around to it on its own accord. The plan is for the MDS to push a caps message informing client B that nlink==0 so that it can prod the VFS into trimming the inode. sage -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
