summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorDavid Teigland <teigland@redhat.com>2023-03-20 12:42:50 -0500
committerDavid Teigland <teigland@redhat.com>2023-03-20 12:42:50 -0500
commit49dbeb9b4566df684164f40ca6ebdab3d6696836 (patch)
treef644d3aa1f9f1ea06f7cb48d6d5842c37a2d12ec
parentddfe839dd1aa42e3d2a15a19eaf86c4b6fc65d19 (diff)
downloadlvm2-49dbeb9b4566df684164f40ca6ebdab3d6696836.tar.gz
tests: nomda-restoremissing
This test was broken by commit 69ea2e3f9 "vg_read: keep MISSING_PV when device with no mda reappears" Change the test to match the new behavior.
-rw-r--r--test/shell/nomda-restoremissing.sh13
1 files changed, 12 insertions, 1 deletions
diff --git a/test/shell/nomda-restoremissing.sh b/test/shell/nomda-restoremissing.sh
index fb3f93053..d9c544bc4 100644
--- a/test/shell/nomda-restoremissing.sh
+++ b/test/shell/nomda-restoremissing.sh
@@ -31,7 +31,18 @@ not vgck $vg 2>&1 | tee log
grep "missing 1 physical volume" log
not lvcreate -aey --type mirror -m 1 -l 1 -n mirror $vg # write operations fail
aux enable_dev "$dev1"
-lvcreate -aey --type mirror -m 1 -l 1 -n mirror $vg # no MDA => automatically restored
+# Old versions would automatically clear MISSING_PV on a PV that had no mda,
+# but this made no sense; the existence of an mda means nothing for the
+# validity of the data on the device. I suspect that at some point in the
+# past, the MISSING_PV flag was used to decide if metadata could be used
+# from the device, so the flag could be cleared on a PV with no mda.
+# These days lvm knows when to ignore outdated metadata.
+# MISSING_PV probably has little to no value for determining valid data either,
+# so it's likely that we'll begin to automatically clear MISSING_PV in the
+# future (but it will have nothing to do with having mdas.)
+not lvcreate -aey --type mirror -m 1 -l 1 -n mirror $vg
+vgextend --restoremissing $vg "$dev1"
+lvcreate -aey --type mirror -m 1 -l 1 -n mirror $vg
vgck $vg
vgremove -ff $vg