Currently running rpki-client -n with an up to date repo results in the
loss of around 25% of ROAs. The reason is that most of apnic fails since
they decided it is a glorious idea to put two rsync repos into one rrdp
repo.
When changing the repo state for the noop case from REPO_DONE to
REPO_FAILED it resulted in a unwanted side-effect. The problem is that a
second rrdp_get() call for such a failed repo fails and so the code falls
back to rsync but there is nothing in the rsync dir.
Just change back the status to REPO_DONE and -n works like before.
--
:wq Claudio
Index: repo.c
===================================================================
RCS file: /cvs/src/usr.sbin/rpki-client/repo.c,v
retrieving revision 1.18
diff -u -p -r1.18 repo.c
--- repo.c 29 Dec 2021 11:35:23 -0000 1.18
+++ repo.c 4 Jan 2022 14:28:41 -0000
@@ -447,7 +447,7 @@ ta_get(struct tal *tal)
tal->uri = NULL;
if (noop) {
- tr->state = REPO_FAILED;
+ tr->state = REPO_DONE;
logx("ta/%s: using cache", tr->descr);
repo_done(tr, 0);
} else {
@@ -512,7 +512,7 @@ rsync_get(const char *uri, int nofetch)
rr->basedir = repo_dir(repo, "rsync", 0);
if (noop || nofetch) {
- rr->state = REPO_FAILED;
+ rr->state = REPO_DONE;
logx("%s: using cache", rr->basedir);
repo_done(rr, 0);
} else {
@@ -582,7 +582,7 @@ rrdp_get(const char *uri, int nofetch)
RB_INIT(&rr->deleted);
if (noop || nofetch) {
- rr->state = REPO_FAILED;
+ rr->state = REPO_DONE;
logx("%s: using cache", rr->notifyuri);
repo_done(rr, 0);
} else {
