2018-09-14 03:58:20
Alright, so I got Arktika to work. Part of the problem was that I had 6GB of chunks allocated to nodes with 8GB of RAM. Those nodes failed with Java Heap overflow, so those nodes would refuse connection. I reallocated the chunks, only 4GB to 8GB nodes and 8GB to 16GB nodes. I started the worker and nodes and in the mean time before all of the nodes were up, the `WARNING: Error: io.grpc.StatusRuntimeException: UNAVAILABLE: io exceptionio.grpc.StatusRuntimeException: UNAVAILABLE: io exception` persisted. When all of the nodes were up, there were no more errors and the arktika miner is reading out hashrates :slightly_smiling_face:
aikida3k
2018-09-14 03:59:15
congratulations :slightly_smiling_face:
mjay
2018-09-14 03:59:23
What is your hashrate with this setup?
mjay
2018-09-14 04:06:18
So my memory server, the worker was at 800K and it dropped to 650K or so and the other nodes are from 50 to 70K with 8-16GB RAM. These are HPProliant DL380 G6 servers. It may not be the most economical; I'll have to watch once this gets on the exchange, but I had these idle after mining a different CPU coin. That coin isn't profitable to mine. I like this coin and am excited about it so I wanted to get something going
aikida3k
2018-09-14 04:07:38
Really grateful for this helpful community: Thanks @mjay @Fireduck @Clueless @Rotonen @SidGrip
aikida3k
2018-09-14 04:12:22
You´re welcome! What is your bottleneck currently, network?
mjay
2018-09-14 04:13:34
800k is at least 77MB/s, close to the limit of Gbit
mjay
2018-09-14 04:14:23
yeah, I would agree it's the network
aikida3k
2018-09-14 04:15:38
what cpu´s are in your servers?
mjay
2018-09-14 04:16:06
dual Xeon X5650
aikida3k
2018-09-14 04:16:26
I had a similar setup running with DL180 G7 servers, slightly faster CPUs (X5670) which maxed out at 900k
mjay
2018-09-14 04:16:32
They benchmarked as the best value for the dollar
aikida3k
2018-09-14 04:17:00
I switched over to DL580 G7 servers after this
mjay
2018-09-14 04:18:26
I like the idea of running a DL580G7 with that E7. I saw above you posted the link for that server on ebay and I contacted him to see if he'll deliver to the US.
aikida3k
2018-09-14 04:18:56
are there no such servers in the us?
mjay
2018-09-14 04:19:27
You could also check for BL680 (same but as blade, they are even cheaper)
mjay
2018-09-14 04:19:40
I also saw a DL580G7 with over 400 GB RAM that was around 1000USD. That was worthwhile for the RAM but the processor wasn't an E7
aikida3k
2018-09-14 04:19:56
Also the Dell RX600
mjay
2018-09-14 04:19:58
hmm. I haven't seen any E7s
aikida3k
2018-09-14 04:20:22
okay. I'll shop around, thanks.
aikida3k
2018-09-14 04:20:29
Or Ibm X3850
mjay
2018-09-14 04:20:34
same hardware different brand
mjay
2018-09-14 04:20:53
ibm is really loud however, same for the blade center
mjay
2018-09-14 04:21:30
place them on wheels and they move just from the fans
mjay
2018-09-14 04:21:42
lol
aikida3k
checksum mismatch
how to solve it?
2018-09-14 08:36:03
what checksum?
mjay
2018-09-14 08:36:42
@mjay don’t discard oracle and fujitsu sparc for snowblossom if LOUD is an option
Rotonen
2018-09-14 08:37:15
my honda civic beats them all
mjay
nvm i solved it
it was just faulty snowblossom release
2018-09-14 08:57:55
care to clarify?
Rotonen
maybe download didnt finish properly
or something
when i unzip it
but its weird though