From 223d0f25cc8e84fa77f11357e53266a28d11a5a0 Mon Sep 17 00:00:00 2001
From: Colin Walters <walters@verbum.org>
Date: Thu, 24 Aug 2023 13:09:05 -0400
Subject: [PATCH] transposefs: Only autosave-xfs for much larger filesystems

The change in https://github.com/coreos/fedora-coreos-config/pull/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.
---
 .../ignition-ostree-transposefs.sh              | 17 ++++++++++++-----
 1 file changed, 12 insertions(+), 5 deletions(-)

diff --git a/overlay.d/05core/usr/lib/dracut/modules.d/40ignition-ostree/ignition-ostree-transposefs.sh b/overlay.d/05core/usr/lib/dracut/modules.d/40ignition-ostree/ignition-ostree-transposefs.sh
index 6f89c4007e..3ba12a6602 100755
--- a/overlay.d/05core/usr/lib/dracut/modules.d/40ignition-ostree/ignition-ostree-transposefs.sh
+++ b/overlay.d/05core/usr/lib/dracut/modules.d/40ignition-ostree/ignition-ostree-transposefs.sh
@@ -139,13 +139,20 @@ should_autosave_rootfs() {
         return
     fi
     local agcount
+    # We want to run xfs_info on the unmounted filesystem, because actually just
+    # even mounting an XFS filesystem that has grown an excessive number of allocation groups
+    # can be very slow.
     eval $(xfs_info "${root_part}" | grep -o 'agcount=[0-9]*')
-    # Semi-arbitrarily chosen: this is roughly ~64G currently (based on initial
-    # ag sizing at build time) which seems like a good rootfs size at which to
+    # This is roughly ~700GiB currently (based on initial ag sizing at build time) which
+    # seems like a good rootfs size at which to
     # discriminate between "throwaway/short-lived systems" and "long-running
-    # workload systems". It's not like XFS performance is way worse at 128.
-    if [ "$agcount" -lt 128 ]; then
-        echo "Filesystem agcount is $agcount; skipping" >&2
+    # workload systems".  Specifically for e.g. OpenShift, it's below the default node
+    # sizes.
+    local threshold
+    threshold=400
+    echo "${root_part} agcount=$agcount" >&2
+    if [ "$agcount" -lt "${threshold}" ]; then
+        echo "agcount=$agcount is lower than threshold=${threshold}"
         echo 0
         return
     fi