transposefs: Only autosave-xfs for much larger filesystems

[cgwalters]

The change in #2320
has been very problematic for OpenShift because our default node
configuration is always over the threshold, and that causes
significant latency on instance provisioning.

Experimentally bumping to 400 allocation groups, which is about 700GiB.
This is comfortably about the default OpenShift node root disk sizes,
and returns us to the prior status quo.

While we're here, rework the logging a bit so that we always
log the agcount for debugging purposes.

---

[cgwalters]

Draft since this could use some more testing

[cgwalters]

Downstream: https://issues.redhat.com/browse/OCPBUGS-16724

[cgwalters]

Here's a script I wrote to analyze xfs AG counts as the system is grown:

#!/bin/bash
set -euo pipefail
initial=$1
shift
grown=$1
shift

runv() {
    echo "$ $@"
    "$@"
} 

tmpd=$(mktemp -d -p /var/tmp)
target="${tmpd}/backing-file"
runv truncate -s "$initial" "$target"
runv mkfs.xfs "${target}"
runv truncate -s "$grown" "$target"
mntpath="${tmpd}/mnt"
mkdir "${mntpath}"
runv mount -o loop -t xfs "${target}" "${mntpath}"
runv xfs_growfs "${mntpath}" 1>/dev/null
runv xfs_info "${mntpath}"
umount "${mntpath}"
rm ${tmpd} -rf

[dustymabe]

Thanks for working on this. A few comments

[dustymabe]

I added the whitespace fixes over in #2566 so you can ignore that CI failure here.

[dustymabe]

I added the whitespace fixes over in #2566 so you can ignore that CI failure here.

I broke the whitespace fixes out into #2570

[dustymabe]

LGTM

[cgwalters]

Blah, missed a case of output redirection. Fixed now and cleaned up.

[dustymabe]

totally optional: this return can be dropped now.

[dustymabe]

LGTM