Hello there,
I am very happy Ztex customer and BTCMiner user. But there are some problems which makes me worried.
Sometimes FPGA boards are really slowing down when main mining pool is unavailable. BTCMiner tries to reconnect thousands of times to main pool and some of boards are idling. Backup pool reporting 30% of hashing power. I could not determine is it a problem with backup pool or BTCMiner trying to reach main pool and wasting time.
Just look at this log:
2012-12-18T20:54:34: Warning: Connection refused: disabling long polling fo 60s
2012-12-18T20:54:34: 001-0: ztex_ufm1_15y1-04A3918D14-2: Error: Connection refused: Disabling URL http://localhost:4990 for 60s
2012-12-18T20:54:34: 001-0: ztex_ufm1_15y1-04A3465EA1-1: Error: Connection refused: Disabling URL http://localhost:4990 for 60s
2012-12-18T20:54:36: 001-0: ztex_ufm1_15y1-04A3463804-2: Error: Connection refused: Disabling URL http://localhost:4990 for 60s
(...)
every single chip on 1.15y board is reporting that, spamming logfile.
And then again, every minute:
2012-12-18T20:58:34: Warning: Connection refused: disabling long polling fo 60s2012-12-18T20:57:47: 001-0: ztex_ufm1_15y1-04A39E4007-4: Error: Connection refused: Disabling URL http://localhost:4990 for 60s
2012-12-18T20:57:49: 001-0: ztex_ufm1_15y1-04A3918D14-2: Error: Connection refused: Disabling URL http://localhost:4990 for 60s
2012-12-18T20:57:49: 001-0: ztex_ufm1_15y1-04A3919EA0-4: Error: Connection refused: Disabling URL http://localhost:4990 for 60s
(...)
And continues. Why is LP disabled when main pool is unavailable? LP could be reached from backup pool.
My suggestions:
1. Add an option to choose, for how long pool should be disabled when unavailable. 60sec is just too short. When main pool fails, I would like to disable it for 10 minutes instead, to make sure it have time to settle down, whatever happening with it.
2. Disable spamming logfile from every single chip, when main pool is unavailable (when running in cluster mode). Only one message like:
Error: Connection refused: Disabling URL xxx for xxx sec
should be enough.
3. Enable logging of backup pool connections - when main pool refusing connections, BTCMiner should write in log file something like:
Error: Connection refused: Disabling URL http://localhost:4990 for 60s
Using LongPolling URL http://backup_pool:8332/lp
EDIT: I noticed that this message "Using LongPooling ..." with backup pool address appears only when main pool is down during miner startup. Later, when pool goes down, we don't know where miner is connected, there is no message about it.
4. When main pool is down, pressing "i" to see status doesn't work, it's lagging for couple of minutes to get output.
What you guys think?