Skip to content

fix: OIDC token refresh recovery after network disruption #991

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
dmuensterer wants to merge 7 commits into
base: main
Choose a base branch
from
+75 −28
Open

fix: OIDC token refresh recovery after network disruption #991

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
2aeab5d
fix: OIDC token refresh recovery after network disruption
dmuensterer Feb 11, 2026
e63e1d9
fix: escalate to full re-auth after sustained OIDC refresh failure
dmuensterer Feb 11, 2026
3e4e554
Use exponential backoff for OIDC refresh retries.
dmuensterer Feb 11, 2026
53c9e00
Consolidate refresh failure counters and remove 3-failure reset
dmuensterer Feb 12, 2026
b0f3edb
Add NULL checks for redirect parsing in OIDC auth flow
dmuensterer Feb 12, 2026
8060597
Merge branch 'main' into fix/oidc-reconnect-refresh
ekoby Feb 13, 2026
fbee24d
Merge branch 'main' into fix/oidc-reconnect-refresh
ekoby Feb 23, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions inc_internal/oidc.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -60,6 +60,7 @@ struct oidc_client_s {
bool need_refresh;
struct auth_req *request;
tlsuv_http_req_t *refresh_req;
int refresh_failures;
Copy link
Member

@ekoby ekoby Feb 11, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think you only need one counter. I see they are set and incremented at the same time

Copy link
Author

@dmuensterer dmuensterer Feb 12, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yup, fixed!

};

// init
Expand Down
45 changes: 24 additions & 21 deletions library/oidc.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -102,6 +102,7 @@ int oidc_client_init(uv_loop_t *loop, oidc_client_t *clt,
return rc;
}
tlsuv_http_set_ssl(&clt->http, tls);
tlsuv_http_connect_timeout(&clt->http, 15000);

clt->timer = calloc(1, sizeof(*clt->timer));
uv_timer_init(loop, clt->timer);
Expand Down Expand Up @@ -570,12 +571,7 @@ static void oidc_client_set_tokens(oidc_client_t *clt, json_object *tok_json) {
}
if (ttl) {
int32_t t = json_object_get_int(ttl);
if (t <= 60) {
OIDC_LOG(WARN, "token lifetime is too short[%d seconds]. this may cause problems", t);
t = t / 2;
} else {
t = t - 30; // refresh 30 seconds before expiry
}
t = t / 2; // refresh at half the token lifetime
OIDC_LOG(DEBUG, "scheduling token refresh in %d seconds", t);
uv_timer_start(clt->timer, refresh_time_cb, t * 1000, 0);
}
Expand All @@ -588,33 +584,40 @@ static void refresh_cb(tlsuv_http_resp_t *http_resp, const char *err, json_objec

if (http_resp->code == 200 && resp != NULL) {
OIDC_LOG(DEBUG, "token refresh success");
clt->refresh_failures = 0;
oidc_client_set_tokens(clt, resp);
return;
}

if (http_resp->code >= 0 || http_resp->code == UV_EOF) {
// controller may abruptly terminate shutdown connection (EOF) if auth has failed
OIDC_LOG(WARN, "OIDC token refresh failed: %d %s [%s]",
http_resp->code, http_resp->status, err);
if (resp) {
OIDC_LOG(WARN, "response: %s", json_object_get_string(resp));
}
json_object_put(clt->tokens);
clt->tokens = NULL;

oidc_client_start(clt, clt->token_cb);
}

if (http_resp->code == UV_ECANCELED) {
OIDC_LOG(DEBUG, "OIDC token refresh was canceled");
return;
}

if (http_resp->code < 0) { // connection failure, try another refresh
OIDC_LOG(WARN, "OIDC token refresh failed (trying again): %d/%s", http_resp->code, err);
if (http_resp->code < 0) {
clt->refresh_failures++;
OIDC_LOG(WARN, "OIDC token refresh failed (%d/%s), attempt %d",
http_resp->code, err, clt->refresh_failures);
if (clt->refresh_failures >= 3) {
Copy link
Member

@ekoby ekoby Feb 11, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

this should be removed now, right?

OIDC_LOG(WARN, "resetting OIDC connection after %d consecutive failures", clt->refresh_failures);
tlsuv_http_cancel_all(&clt->http);
clt->refresh_failures = 0;
}
uv_timer_start(clt->timer, refresh_time_cb, 5 * 1000, 0);
return;
}

// http_resp->code >= 0 but not 200: server-side rejection (e.g. EOF, 401, etc.)
Copy link
Member

@ekoby ekoby Feb 11, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

EOF(UV_EOF) is negative, so it will be covered in the if(http_resp->code < 0) clause above

Copy link
Author

@dmuensterer dmuensterer Feb 11, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Ofc you are right, I removed it. I fixed the comment on the remaining branch to say "e.g. 401, 403" instead of "e.g. EOF, 401" since EOF won't reach this clause here?

OIDC_LOG(WARN, "OIDC token refresh failed: %d %s [%s]",
http_resp->code, http_resp->status, err);
if (resp) {
OIDC_LOG(WARN, "response: %s", json_object_get_string(resp));
}
clt->refresh_failures = 0;
json_object_put(clt->tokens);
clt->tokens = NULL;

oidc_client_start(clt, clt->token_cb);
}

static const char* get_basic_auth_header(const char *client_id) {
Expand Down