1

I am not new to linux, but I am new to the underlying guts of it, and I don't know anything about LVM or device-mapper. I'm trying to troubleshoot a livecd mounting its live-rw (dm-0) file system as read only during operation, and it's been frustrating to put it mildly.

I've discovered that dmsetup status will display some numbers - I think snapshot size, but I'm not sure:

[root@localhost ~]# dmsetup status
live-osimg-min: 0 4194304 snapshot 1272/1272 16
live-rw: 0 4194304 snapshot 7568/1048576 40

I've found that the far right numbers, the "7568/1048576" seem to determine the length of time my livecd has to live before it gets mounted read-only and dies. These numbers always start at the same amount and tick up as more activity is performed on the file system - regardless of the root partition's size. Commands like "yum update" are particularly harsh on this number.

I can't find any documentation online that labels the fields and describes what they are - in classic linux fashion, they're just some floating integers on a black background with absolutely no explanation, as if to say "good luck!".

Does anyone know what these numbers represent, and how I can increase the "cap" from 1048576 to something much larger when building a CentOS 6.6 livecd?

1 Answer 1

1

The Linux kernel documentation of the snapshot DM target contains the answer to your first question:

The snapshot-merge and snapshot status lines end with: <sectors_allocated>/<total_sectors> <metadata_sectors>

The first two numbers (0 and 4194304 in your case) define the start and the length of the mapping in 512-byte sectors, but this is not specific to the snapshot target (it isn't strictly status info, but more like table info). Read the other files in the device-mapper documentation directory if you are interested.

The snapshot size ("cap") is determined by the size of the copy-on-write (COW) device given to the snapshot target. Check the target parameters via dmsetup table. Their documentation is in the same file, in short:

snapshot <origin> <COW device> <persistent?> <chunksize>

What your CentOS 6.6 LiveCD uses as a COW device and how to influence it, I don't know.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .