use diesel::{sql_query, PgConnection, RunQueryDsl};
use lemmy_api_common::context::LemmyContext;
use lemmy_db_schema::{
- schema::{activity, comment, community_person_ban, instance, person, post},
+ schema::{activity, captcha_answer, comment, community_person_ban, instance, person, post},
source::instance::{Instance, InstanceForm},
utils::{naive_now, DELETED_REPLACEMENT_TEXT},
};
use lemmy_routes::nodeinfo::NodeInfo;
-use lemmy_utils::{error::LemmyError, REQWEST_TIMEOUT};
+use lemmy_utils::{
+ error::{LemmyError, LemmyResult},
+ REQWEST_TIMEOUT,
+};
use reqwest::blocking::Client;
use std::{thread, time::Duration};
-use tracing::{error, info};
+use tracing::{error, info, warn};
/// Schedules various cleanup tasks for lemmy in a background thread
pub fn setup(
update_hot_ranks(&mut conn, true);
});
+ // Delete any captcha answers older than ten minutes, every ten minutes
+ let url = db_url.clone();
+ scheduler.every(CTimeUnits::minutes(10)).run(move || {
+ let mut conn = PgConnection::establish(&url).expect("could not establish connection");
+ delete_expired_captcha_answers(&mut conn);
+ });
+
// Clear old activities every week
let url = db_url.clone();
scheduler.every(CTimeUnits::weeks(1)).run(move || {
// Update the Instance Software
scheduler.every(CTimeUnits::days(1)).run(move || {
let mut conn = PgConnection::establish(&db_url).expect("could not establish connection");
- update_instance_software(&mut conn, &user_agent);
+ update_instance_software(&mut conn, &user_agent)
+ .map_err(|e| warn!("Failed to update instance software: {e}"))
+ .ok();
});
// Manually run the scheduler in an event loop
);
}
+fn delete_expired_captcha_answers(conn: &mut PgConnection) {
+ match diesel::delete(
+ captcha_answer::table.filter(captcha_answer::published.lt(now - IntervalDsl::minutes(10))),
+ )
+ .execute(conn)
+ {
+ Ok(_) => {
+ info!("Done.");
+ }
+ Err(e) => {
+ error!("Failed to clear old captcha answers: {}", e)
+ }
+ }
+}
+
/// Clear old activities (this table gets very large)
fn clear_old_activities(conn: &mut PgConnection) {
info!("Clearing old activities...");
for i in &intervals {
let update_site_stmt = format!(
- "update site_aggregates set users_active_{} = (select * from site_aggregates_activity('{}'))",
+ "update site_aggregates set users_active_{} = (select * from site_aggregates_activity('{}')) where site_id = 1",
i.1, i.0
);
match sql_query(update_site_stmt).execute(conn) {
}
/// Updates the instance software and version
-fn update_instance_software(conn: &mut PgConnection, user_agent: &str) {
+///
+/// TODO: this should be async
+/// TODO: if instance has been dead for a long time, it should be checked less frequently
+fn update_instance_software(conn: &mut PgConnection, user_agent: &str) -> LemmyResult<()> {
info!("Updating instances software and versions...");
- let client = match Client::builder()
+ let client = Client::builder()
.user_agent(user_agent)
.timeout(REQWEST_TIMEOUT)
- .build()
- {
- Ok(client) => client,
- Err(e) => {
- error!("Failed to build reqwest client: {}", e);
- return;
- }
- };
+ .build()?;
- let instances = match instance::table.get_results::<Instance>(conn) {
- Ok(instances) => instances,
- Err(e) => {
- error!("Failed to get instances: {}", e);
- return;
- }
- };
+ let instances = instance::table.get_results::<Instance>(conn)?;
for instance in instances {
let node_info_url = format!("https://{}/nodeinfo/2.0.json", instance.domain);
- // Skip it if it can't connect
- let res = client
- .get(&node_info_url)
- .send()
- .ok()
- .and_then(|t| t.json::<NodeInfo>().ok());
-
- if let Some(node_info) = res {
- let software = node_info.software.as_ref();
- let form = InstanceForm::builder()
- .domain(instance.domain)
- .software(software.and_then(|s| s.name.clone()))
- .version(software.and_then(|s| s.version.clone()))
- .updated(Some(naive_now()))
- .build();
-
- match diesel::update(instance::table.find(instance.id))
- .set(form)
- .execute(conn)
- {
- Ok(_) => {
- info!("Done.");
+ // The `updated` column is used to check if instances are alive. If it is more than three days
+ // in the past, no outgoing activities will be sent to that instance. However not every
+ // Fediverse instance has a valid Nodeinfo endpoint (its not required for Activitypub). That's
+ // why we always need to mark instances as updated if they are alive.
+ let default_form = InstanceForm::builder()
+ .domain(instance.domain.clone())
+ .updated(Some(naive_now()))
+ .build();
+ let form = match client.get(&node_info_url).send() {
+ Ok(res) if res.status().is_client_error() => {
+ // Instance doesnt have nodeinfo but sent a response, consider it alive
+ Some(default_form)
+ }
+ Ok(res) => match res.json::<NodeInfo>() {
+ Ok(node_info) => {
+ // Instance sent valid nodeinfo, write it to db
+ Some(
+ InstanceForm::builder()
+ .domain(instance.domain)
+ .updated(Some(naive_now()))
+ .software(node_info.software.and_then(|s| s.name))
+ .version(node_info.version.clone())
+ .build(),
+ )
}
- Err(e) => {
- error!("Failed to update site instance software: {}", e);
- return;
+ Err(_) => {
+ // No valid nodeinfo but valid HTTP response, consider instance alive
+ Some(default_form)
}
+ },
+ Err(_) => {
+ // dead instance, do nothing
+ None
}
+ };
+ if let Some(form) = form {
+ diesel::update(instance::table.find(instance.id))
+ .set(form)
+ .execute(conn)?;
}
}
+ info!("Finished updating instances software and versions...");
+ Ok(())
}
#[cfg(test)]
projects / lemmy.git / blobdiff