Browse Source

Detect when the primary pool is lagging and start queueing requests on backup pools if possible before needing to roll work.

nfactor-troky
Con Kolivas 13 years ago
parent
commit
508c5c4b12
  1. 29
      main.c

29
main.c

@ -87,6 +87,7 @@ struct workio_cmd { @@ -87,6 +87,7 @@ struct workio_cmd {
union {
struct work *work;
} u;
bool lagging;
};
enum sha256_algos {
@ -930,14 +931,14 @@ static const char *rpc_req = @@ -930,14 +931,14 @@ static const char *rpc_req =
"{\"method\": \"getwork\", \"params\": [], \"id\":0}\r\n";
/* Select any active pool in a rotating fashion when loadbalance is chosen */
static inline struct pool *select_pool(void)
static inline struct pool *select_pool(bool lagging)
{
static int rotating_pool;
static int rotating_pool = 0;
struct pool *pool, *cp;
cp = current_pool();
if (pool_strategy != POOL_LOADBALANCE)
if (pool_strategy != POOL_LOADBALANCE && !lagging)
pool = cp;
else
pool = NULL;
@ -954,18 +955,20 @@ static inline struct pool *select_pool(void) @@ -954,18 +955,20 @@ static inline struct pool *select_pool(void)
return pool;
}
static bool get_upstream_work(struct work *work)
static bool get_upstream_work(struct work *work, bool lagging)
{
struct pool *pool = select_pool();
struct pool *pool;
json_t *val;
bool rc = false;
CURL *curl = curl_easy_init();
CURL *curl;
curl = curl_easy_init();
if (unlikely(!curl)) {
applog(LOG_ERR, "CURL initialisation failed");
return rc;
}
pool = select_pool(lagging);
val = json_rpc_call(curl, pool->rpc_url, pool->rpc_userpass, rpc_req,
want_longpoll, false, pool);
if (unlikely(!val)) {
@ -1099,7 +1102,7 @@ static void *get_work_thread(void *userdata) @@ -1099,7 +1102,7 @@ static void *get_work_thread(void *userdata)
}
/* obtain new work from bitcoin via JSON-RPC */
while (!get_upstream_work(ret_work)) {
while (!get_upstream_work(ret_work, wc->lagging)) {
if (unlikely((opt_retries >= 0) && (++failures > opt_retries))) {
applog(LOG_ERR, "json_rpc_call failed, terminating workio thread");
free(ret_work);
@ -1987,10 +1990,14 @@ static bool queue_request(void) @@ -1987,10 +1990,14 @@ static bool queue_request(void)
{
int maxq = opt_queue + mining_threads;
struct workio_cmd *wc;
int rq, rs;
rq = requests_queued();
rs = real_staged();
/* If we've been generating lots of local work we may already have
* enough in the queue */
if (requests_queued() >= maxq || real_staged() >= maxq)
if (rq >= maxq || rs >= maxq)
return true;
/* fill out work request message */
@ -2004,6 +2011,12 @@ static bool queue_request(void) @@ -2004,6 +2011,12 @@ static bool queue_request(void)
/* The get work does not belong to any thread */
wc->thr = NULL;
/* If we've queued more than 2/3 of the maximum and still have no
* staged work, consider the system lagging and allow work to be
* gathered from another pool if possible */
if (rq > (maxq * 2 / 3) && !rs)
wc->lagging = true;
/* send work request to workio thread */
if (unlikely(!tq_push(thr_info[work_thr_id].q, wc))) {
applog(LOG_ERR, "Failed to tq_push in queue_request");

Loading…
Cancel
Save