1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
|
summary: Ensure catalog update works
details: |
Verify that catalog refresh happens on snapd startup. Also check
unnecessary catalog body data is not logged.
environment:
# on core systems, the catalog update is not performed when memory limit is
# in place
SNAPD_NO_MEMORY_LIMIT: 1
prepare: |
echo "Remove cached catalog files from disk"
rm -f /var/cache/snapd/names
rm -f /var/cache/snapd/sections
execute: |
echo "We count how many catalog refreshes are logged before starting snapd"
refreshes_before="$("$TESTSTOOLS"/journal-state get-log -u snapd | grep -c 'Catalog refresh' || true)"
systemctl restart snapd.{socket,service}
echo "Ensure that catalog refresh happens on startup"
for _ in $(seq 60); do
refreshes_after="$("$TESTSTOOLS"/journal-state get-log -u snapd | grep -c 'Catalog refresh' || true)"
if [ "$refreshes_after" -gt "$refreshes_before" ]; then
break
fi
sleep 1
done
echo "Ensure the current number of refreshes is greater than before restarting snapd"
[ "$refreshes_after" -gt "$refreshes_before" ]
echo "Ensure that we don't log all catalog body data"
if "$TESTSTOOLS"/journal-state get-log -u snapd | MATCH "Tools for testing the snapd application"; then
echo "Catalog update is doing verbose http logging (it should not)."
exit 1
fi
|
