[pdm-devel] [PATCH datacenter-manager 2/9] server: task cache: skip remotes with errors on fetch

[Thomas Lamprecht]

Am 20.01.25 um 10:29 schrieb Dominik Csapak:
> instead of failing the whole cache retrieval, just skip the remote.
> 
> Signed-off-by: Dominik Csapak <d.csapak at proxmox.com>
> ---
>  server/src/task_cache.rs | 8 +++++++-
>  1 file changed, 7 insertions(+), 1 deletion(-)
> 
> diff --git a/server/src/task_cache.rs b/server/src/task_cache.rs
> index 3ef24b1..210aa5f 100644
> --- a/server/src/task_cache.rs
> +++ b/server/src/task_cache.rs
> @@ -42,7 +42,13 @@ pub async fn get_tasks(max_age: i64) -> Result<Vec<TaskListItem>, Error> {
>              // Data in cache is recent enough and has not been invalidated.
>              all_tasks.extend(tasks);
>          } else {
> -            let tasks = fetch_tasks(remote).await?;
> +            let tasks = match fetch_tasks(remote).await {
> +                Ok(tasks) => tasks,
> +                Err(err) => {
> +                    // ignore errors for not reachable remotes

Even if it's now the case, it's not hard-coded for the future that an
error here will always stem from an unreachable remote; at least
logging that as warning might be good to allow debugging odd things
in a setup.

> +                    continue;
> +                }
> +            };
>              cache.set_tasks(remote_name.as_str(), tasks.clone(), now);
>  
>              all_tasks.extend(tasks);