summaryrefslogtreecommitdiff
path: root/doc/forum
diff options
context:
space:
mode:
authorJoey Hess2019-07-10 22:18:23 -0400
committerJoey Hess2019-07-10 22:18:23 -0400
commit31a8f5390d01788f794055f1fdf39326d597d48d (patch)
treed495ec251a4f728058962f7528a3c6377d2d8993 /doc/forum
parent4064ab26f96709010583840ceea9f68fa08cf495 (diff)
seeking advice from DDs
Diffstat (limited to 'doc/forum')
-rw-r--r--doc/forum/apt_releaseinfo.mdwn17
1 files changed, 17 insertions, 0 deletions
diff --git a/doc/forum/apt_releaseinfo.mdwn b/doc/forum/apt_releaseinfo.mdwn
new file mode 100644
index 00000000..f817da69
--- /dev/null
+++ b/doc/forum/apt_releaseinfo.mdwn
@@ -0,0 +1,17 @@
+Since apt-get has changed behavior and refuses to update if the code name
+in a Release file has changed from the one it saw before, propellor systems
+that are tracking Unstable or Testing need a manual intervention after each
+major stable release (`apt-get update --allow-releaseinfo-change`).
+
+(It seems that when unattended-updates is available, it does automatically
+accept the change.)
+
+I am seeking advice about what propellor should do about this. Should
+Apt.update include --allow-releaseinfo-change when the host is following a
+rolling suite?
+
+Or does doing so defeat whatever presumably good rationalle there is for
+making this a manual action? Or perhaps there is no really good reason apt
+does this when following a folling suite?
+
+--[[Joey]]