Thanks Alex. Very helpful! I’m getting the equipment this week so will set it up and do some benchmarking and come up with a plan. May not be able to use all the disks or space. Will see. Thanks again.
Watched your 1:N videos and read the articles. Had three nodes running and tried to merge those three into two nodes (early and late). Never used team24 so set that up and now have two nodes with two post servers running to early and one running to late. Post data is on SSD. Two are 2TB and last 1TB. Problem is I don’t know if everything is working correctly again. Grafana shows the late node not activating in time (red) and not sure if that is normal. Early node is green. team24 so far only shows one post of 2 early nodes as seen as far as I understand the page. I’ve checked and rechecked configuration and everything seems correct. So I may have to wait until end of this epoch and start of next to see if I’m smeshing again. Your articles and videos are very instructive! Thanks
Thanks Alex. I think I’ve tried this but couldn’t create second container because it could not find spacenet or simple-xxx_spacenet (not home don’t remember entire name), saying it was either not external or couldn’t find it. Anyway I’ll read over this and look at the link and try again once I get home. Thanks. Rand
Alternately you can just create a network in the terminal. The command is in the linked article (just rename it to 'spacemesh-network' or something. Then you should have no issues importing it into you stack file
Alex, I found your “How to increase reward potential” for SpaceMesh and that answered a lot of questions which I was raising in my second post about the DS4246. So I’m continue to study that post and try to figure out the DS4246 and configuring SpaceMesh to use some of it. Thanks
Another quick question if you don't mind. Based on this video I purchased a DS4246 12 GB/s 24 drive Netapp with dual IOM-12 . It includes 24 cheap 4TB drives (probably 3 GB/s because of price but don't know). Also have purchased a LSI 9300-8e T93TG which is supposed to support dual channel. I also understand that it is not suitable for Autonomy (which I'm also running). Once I get it running in Linux, what do you think would be the best way to configure for Spacemesh? Can use LVM to make a file system as large as 94TB or something like that, or smaller sized file systems. What do you think is the best way to use this? Maybe 2 45 TB Subspace nodes, one early and one late? Any quick suggestions. Will be running on an Intel i9 pc with 128 GB memory and 4090 GPU.
This is a good question - I have not done anything with Spacemesh in over 6 months so I am not extremely confident on answering this. You definitely want to go as big as you can and still able to finish proofing in time. That will depend mostly on the real life read speed of the disks. You would need some sort of raid if you want to get the speed required to do something like 45TB in a single 24 hour cycle gap. It should be possible but I do not have experience. I chose to go the route of splitting 16 and 18TB disks and having a single plot per disk. I'm not sure if that is the best way to do it these days though, a few large plots might be the best option.
Very helpful Alex. I do need help with something if it is easy. I followed your original directions for setting up Portainer, Grafana, etc for Spacemesh. So all of my Spacemesh containers are running inside that base stack, network 172.0.18.x. I would rather set up a separate stack just for the spacemesh containers but use the network that was setup in the base stack. Is that possible or do I need to setup a stack with a different network address, e.g. 172.0.19.x? I've tried all sorts of things without success. Is there a way to share the 172.0.18.x address among multiple stacks or does each need its on network? Thanks. Rand
You can get the network name from the "Networks" section, just find the one that matches 172.0.18.1. Of course just make sure when deploying the stack that there is no IP conflicts, but it should all be on the same network.
Thanks Alex. Very helpful! I’m getting the equipment this week so will set it up and do some benchmarking and come up with a plan. May not be able to use all the disks or space. Will see. Thanks again.
Watched your 1:N videos and read the articles. Had three nodes running and tried to merge those three into two nodes (early and late). Never used team24 so set that up and now have two nodes with two post servers running to early and one running to late. Post data is on SSD. Two are 2TB and last 1TB. Problem is I don’t know if everything is working correctly again. Grafana shows the late node not activating in time (red) and not sure if that is normal. Early node is green. team24 so far only shows one post of 2 early nodes as seen as far as I understand the page. I’ve checked and rechecked configuration and everything seems correct. So I may have to wait until end of this epoch and start of next to see if I’m smeshing again. Your articles and videos are very instructive! Thanks
Thanks Alex. I think I’ve tried this but couldn’t create second container because it could not find spacenet or simple-xxx_spacenet (not home don’t remember entire name), saying it was either not external or couldn’t find it. Anyway I’ll read over this and look at the link and try again once I get home. Thanks. Rand
Alternately you can just create a network in the terminal. The command is in the linked article (just rename it to 'spacemesh-network' or something. Then you should have no issues importing it into you stack file
Alex, I found your “How to increase reward potential” for SpaceMesh and that answered a lot of questions which I was raising in my second post about the DS4246. So I’m continue to study that post and try to figure out the DS4246 and configuring SpaceMesh to use some of it. Thanks
Another quick question if you don't mind. Based on this video I purchased a DS4246 12 GB/s 24 drive Netapp with dual IOM-12 . It includes 24 cheap 4TB drives (probably 3 GB/s because of price but don't know). Also have purchased a LSI 9300-8e T93TG which is supposed to support dual channel. I also understand that it is not suitable for Autonomy (which I'm also running). Once I get it running in Linux, what do you think would be the best way to configure for Spacemesh? Can use LVM to make a file system as large as 94TB or something like that, or smaller sized file systems. What do you think is the best way to use this? Maybe 2 45 TB Subspace nodes, one early and one late? Any quick suggestions. Will be running on an Intel i9 pc with 128 GB memory and 4090 GPU.
Thanks Rand
This is a good question - I have not done anything with Spacemesh in over 6 months so I am not extremely confident on answering this. You definitely want to go as big as you can and still able to finish proofing in time. That will depend mostly on the real life read speed of the disks. You would need some sort of raid if you want to get the speed required to do something like 45TB in a single 24 hour cycle gap. It should be possible but I do not have experience. I chose to go the route of splitting 16 and 18TB disks and having a single plot per disk. I'm not sure if that is the best way to do it these days though, a few large plots might be the best option.
Sorry I can't be more help
Very helpful Alex. I do need help with something if it is easy. I followed your original directions for setting up Portainer, Grafana, etc for Spacemesh. So all of my Spacemesh containers are running inside that base stack, network 172.0.18.x. I would rather set up a separate stack just for the spacemesh containers but use the network that was setup in the base stack. Is that possible or do I need to setup a stack with a different network address, e.g. 172.0.19.x? I've tried all sorts of things without success. Is there a way to share the 172.0.18.x address among multiple stacks or does each need its on network? Thanks. Rand
Hey Rand, yes you can share the network by importing it into your stack file. You can do it like this:
networks:
NETWORK_NAME_HERE:
external: true
The formatting doesn't present well in this comment - but if you check this guide it has the code at the bottom of the stackfile: https://hakedev.substack.com/p/ultimate-autonomys-guide-pt-3
You can get the network name from the "Networks" section, just find the one that matches 172.0.18.1. Of course just make sure when deploying the stack that there is no IP conflicts, but it should all be on the same network.