From acd309acd666a0d065a3d8cd91582e964e20184a Mon Sep 17 00:00:00 2001 From: Walter Doekes Date: Tue, 11 May 2021 12:10:27 +0200 Subject: [PATCH] presence: Fix leaking of activewatchers in database (with clustering) If you're using clustering, without tags (or with tags, but without fallback2db), you would end up with a lot of active watchers in the database that never got cleaned up. Before 929ab4d43c0ad1878d60d3ef27f2a71f905b80f0: - all stale activewatcher records in the db got purged. After: - if you're not clustering, all stale active watcher got purged; - if you are, only those with the sharing_tag set. However, the sharing tag is not necessarily set: - it is an optional argument to handle_subscribe; - and setting it in handle_subscribe does not write to the database because of missing code in the fallback2db==0 bit; - and even it it were set, then adding the argument to handle_subscribe does not "activate" the sharing tag. (Also interesting to know: a 408 or 481 after the this-subscription-is-expired NOTIFY _would_ cause the individual record to get deleted. But any other response, including 200, would leave the record to get sorted by the periodic purge.) This changeset reverts parts of the aforementioned commit by always purging stale records if the sharing_tag is NULL. Thus restoring behaviour to pre-3.1.0 and pre-2.4.8. --- modules/presence/subscribe.c | 31 +++++++++++++++---------------- 1 file changed, 15 insertions(+), 16 deletions(-) diff --git a/modules/presence/subscribe.c b/modules/presence/subscribe.c index d8b91a3edee..eb547404082 100644 --- a/modules/presence/subscribe.c +++ b/modules/presence/subscribe.c @@ -1357,7 +1357,7 @@ static inline int is_shtag_active( str *my_tag, str **active_tags) void update_db_subs(db_con_t *db,db_func_t *dbf, shtable_t hash_table, int htable_size, int no_lock, handle_expired_func_t handle_expired_func) { - static db_ps_t my_ps_delete = NULL; + static db_ps_t my_ps_delete = NULL, my_ps_delete_null = NULL; static db_ps_t my_ps_update = NULL, my_ps_insert = NULL; db_key_t query_cols[22], update_cols[8]; db_val_t query_vals[22], update_vals[8]; @@ -1660,33 +1660,33 @@ void update_db_subs(db_con_t *db,db_func_t *dbf, shtable_t hash_table, lock_release(&hash_table[i].lock); } - /* now that all records were updated, delete whatever + /* now that all records were updated, delete whatever was still left as expired */ - update_cols[0]= &str_expires_col; + update_cols[0] = &str_expires_col; update_vals[0].type = DB_INT; update_vals[0].nul = 0; update_vals[0].val.int_val = (int)time(NULL); update_ops[0] = OP_LT; - CON_SET_CURR_PS(db, &my_ps_delete); + update_cols[1] = &str_sharing_tag_col; + update_vals[1].nul = 1; + update_ops[1] = OP_IS_NULL; + if (dbf->use_table(db, &active_watchers_table) < 0) { LM_ERR("deleting expired information from database\n"); CON_RESET_CURR_PS(db); return; } - if (sh_tags==NULL) { - - /* no clustering, simply delete all expired subs */ - LM_DBG("delete all expired subscriptions\n"); + /* no clustering, simply delete all expired subs with NULL sh tags */ + LM_DBG("delete all expired subscriptions\n"); - if (dbf->delete(db, update_cols, update_ops, update_vals, 1) < 0) - LM_ERR("deleting expired information from database\n"); - - } else { + CON_SET_CURR_PS(db, &my_ps_delete_null); + if (dbf->delete(db, update_cols, update_ops, update_vals, 2) < 0) + LM_ERR("deleting expired information from database\n"); + if (sh_tags != NULL) { /* clustering, delete only expired subs with active sh tags */ - update_cols[1]= &str_sharing_tag_col; update_vals[1].type = DB_STR; update_vals[1].nul = 0; update_ops[1] = OP_EQ; @@ -1697,13 +1697,12 @@ void update_db_subs(db_con_t *db,db_func_t *dbf, shtable_t hash_table, sh_tags[i]->len, sh_tags[i]->s); update_vals[1].val.str_val = *sh_tags[i]; + + CON_SET_CURR_PS(db, &my_ps_delete); if (dbf->delete(db, update_cols, update_ops, update_vals, 2) < 0) LM_ERR("deleting expired information from database\n"); i++; } - - if (i == 0) - CON_RESET_CURR_PS(db); } return;