Hi all, i have 3 vg's, each with a single lvol. On just one of the 260GB-sized vg's i'm missing almost 17GB! This doesn't happen on the other vg's where everything seems normal. The output of bdf -i on that mountpoint shows that the total number of inodes is huge (64786916) when compared to the number of inodes of the other lvols(1988864).
Hi, let me explain better. I'm sure that it's not a process that keeps a file locked. I even unmounted the mountpoint and remounted. The difference between actual lvol size and used space + free space is 17GB. I'm attaching an output of bdf -i of the mountpoint in order to make it easier to explain. Notice the insanely huge number of inodes.
Which 17 GB you are talking about. The file system has everything free and has only around 80K used and that is understood to be used by file system structures.
Is there any difference in LVOL size and file system size. If yes then somebody extended the LVOL after creating the file system and you can extend file system as well using extendfs.
Can you also attach #lvdisplay -v /dev/vgoracle3/lvol01
Was this lvol formatted as a VxFS filesystem of HFS? HFS must pre-create all the inodes when making the filesystem for the first time. Also, has this lcol been used in the past to store thousands of files that were later removed? Do you have the option to remove the files on this lvol and reformat the filesystem with newfs?
The lvol was formatted with vxfs from the beginning. I have already destroyed and recreated the vg 3 times, changing the PE size. What i've noticed is that there is a difference between the output of bdf and df -k, even though i've unmounted, vgexported and recreated the vg! Starting a restore on the mountpoint seems to have woken bdf since its starting to report much more sane sizes. However, since this seems to be a rather complicated case, i've logged a call with HP.
