Hi SPO,
I’ll ask the exact same question i asked for Mary fork.
Do 8 Gb RAM is enought for Alonzo and smart contracts on september 12 ? I think the awnser is no, big NO.
Do you think IOHK should update the minimum requirements page ?
If they do, that’s mean all nodes on raspberry pie will be obsolete. There’s a lot of raspberry pie on Cardano network.
Cloud base VM with 16gb RAM are expensive and it’s not all SPO that can buy good hardware for bare-metal node. There is a fondamantal reflexion to have here.
Six months ago, i started with only 4 Gb RAM, since then i upgraded nodes 2 times. 4 Gb > 8 Gb > 16 Gb. I hope it’s gona be enought to make it to christmass. 
Luc [LvLUp]
IMO 16GB is the minimum you need, even now that is the ideal. You may notice on your server with Shelley it uses up to 95% of 8GB regularly, making it difficult to run additional applications including monitoring services.
Is anyone running a stake pool with RPi 4 8gb on alonzo testnet?
For the beginning or transition you can probably solve this with the swap file.
I get that everyone likes to run Cardano on an inexpensive RPi but I don’t think I would bother putting such a heavy hitting software suite on one. For coolness factor, maybe. I just put bought a Dell Optiplex 9020-I7 with 32GB of ram for $200/USD and it’s more than enough for me. It only pulls 18 watts of power while running.
Minimum Stake Pool Hardware Requirements for release 1.29
Processor:** An Intel or AMD x86 processor with two or more cores, at 2GHz or faster
Memory:** 8GB of RAM
Storage:** 50GB of free storage
Could you please explain little bit more. I just started 2 days ago.
What needs to be done to make 8Gb not enough?
I’m worried that I haven’t done enough.
Thank you.
I agree with you, I think 16GB is consider safe and most desireable amount. IMO.
Bigger memory always better.
But as a SPO without any blocks and delegators and with dream to get one someday, want to know reasons (it means the pool does not support itself).
Imho, if server works not 90-100% of its power, owner lost money.
Now we can see, that server still has about 2Gb RAM free.
After Alonzo update, every 15 minutes relay rise RAM from 2.5Gb to 5.2Gb and freezing for 30-40 sec. Not restarted, just freezd, no logs, nothing. And after start again from 2.5Gb RAM.
After when I updated relay RAM from 8 to 12Gb, it looks good, without freezing, but still use not more than 6Gb. It’s strange and sad.
Keep in mind pal, cardano-node isn’t the only consumer of memory on your server. Some of this may be other services – specifically drive cache.
Indeed, i thought about it too late… thank you to remind.
Now looks fair, 9Gb
FWIW, I had a local relay at 12GB and another at 16GB. Same host machine. The 12GB would drag at times. 16GB, no problems. Both local relays are now at 16GB. Also, 16GB for BP node, same host, no problems.
8GB, 2 CPU Azure relay has been tweaked for RAM savings - including no monitoring other than atop. It is still struggling (CPU stays at 100%++). Had to reload the blockchain, too. It’s a freebie vm, so…meh.
You can try to low the number of peers connected to your node. Setup topologyUpdater.
More peers you are connected on, more RAM the relay use.
On relays, with minimum 15 peers and Prometheus running, i clearly go above 8Gb.
https://dashboard.lvlup-stakepool.com
I’m running cardano-node on alonzo era just for minting NFT and even I’ve started to notice that my 8 GB VPS is crashing randomly.
The memory usage is through the roof.
May I ask, why 50GB of HD? as far as I read, it only uses 20GB.
By the way, a side question would be, can swap be used to increase the RAM from 8GB to 12GB? If so, it is a matter of knowing what is more expensive, RAM or HD.
