tx pool: fails to relay some txs, and stops trying

[nahuhh]

likely due to frequent disconnects, eventually the node ended up with many txs that are in a "not relayed" state.

a manual relay_tx works, so there is no good reason for these txs to be stuck. I assume the bug is that it stops trying to relay them.

related but not reproduced yet: These were there for over 12hrs, and (strangely) causing double spend errors via rpc but not cli. i could open the wallet in cli and send many txs, but it would fail consistently when attempting the same txs from rpc. I've since flushed i flushed this pool.

probably upstream issue

print_pool_stats

7361 tx(es), 78435327 bytes total (min 6498, max 723486, avg 10655, median 6498)                                   
fees 4.870505640000 (avg 0.000661663583 per tx, 0.000000062095 per byte)                                                                                                                                                              
36 double spends, 78 not relayed, 0 failing, 7128 older than 10 minutes (oldest 2 days ago), estimated 252 block (504 minutes) backlog                                                                                                
   Age      Txes       Bytes                                                                                                                                                                                                          
00:00:00     608    18718051                                                                                                                                                                                                          
00:37:24       6     4340916                                                                                                                                                                                                          
01:14:49     202     7048500                                                                                                                                                                                                          
01:52:13     166     5649540
02:29:38     225     1501271
03:07:02       2       12996
03:44:27       2       12996
04:21:51     337     2189826
04:59:16    5666    36817668
70:44:57     147     2143563

[nahuhh]

monerod print_pool_sh | grep -B6 "relayed: no" will show the non-relayed txs. Running relay_tx <txid> will relay the tx successfully