This is purely out of my own curiosity and isn’t being used for anything by the company or otherwise.
If you are a SNO would you mind telling me how many nodes you operate? (I dont need to know where they are located or what they are running on)
And can you tell me, on average, how much time you think you spend on your nodes per month performing any kind of maintenance? Note, not time setting up new nodes, or interfacing new tools or scripts with existing nodes. I’m primarily interested in how much time is spent with nodes that are completely setup.
I’m hoping to hear most SNOs don’t look at their nodes unless they get a “Your Node has Gone Offline” email. Just turn their payout into cash at the start of every month… but otherwise ignore Storj entirely…
Currently — 7 nodes. 0 time. (after removing outliers like reproducing and reporting bugs couple of times — I did’n have to do that). Nodes don’t require any maintenance. Once setup — they run. If “AllHealthy” flag drops I’ll get an email. Then I’ll investigate. Did not need to do that in the last couple of years — all failures were due to power failure I already knew about through other means.
This format will get you very skewed view, however — a lot of sno who don’t interact with their nodes I imagine won’t be hanging out on this forum either.
Perhaps it is worth making this company wide initiative, and send a proper survey — this can be an important perceived quality metric.
I run 7 nodes, and on average, I spend about an hour a month on analyzing dashboards to spot anything out of the ordinary and checking logs for critical errors. If there’s an issue, it can take a bit more time, but otherwise, it’s pretty low-maintenance once set up.
I currently run 16, 10 get no egress and will be gracefully exited from the 1 satellite they have left when they get too small to care about, and 6 will remain. I stare at graphs way too much, but for the most part the nodes require no active work anymore. So 0 hours.
17 nodes. Zero maintenance time in the recent months. Only when there’s something realy new and I test that setting, I check them from time to time. Otherwise they do their business, I do mine. And this should be the goal… they should not need any maintenance.
Three nodes, average of one hour per month maintenance, primarily Ubuntu updates and review of performance and reviewing forum updates. Last month spent more time than usual as one node was acting unusual and so I started Graceful Exit.
16 nodes
I spend, perhaps, an hour per month. It was more than that during The Days of the Great Ingress when all my machines were suffering but now it’s all calmed down
4 nodes.
1 on Wintel, 3 on Syno NAS, one node per HDD, only 1 HDD is dedicated to storj. Time would be zero, as I don’t count reading this forum and staring on dashboard as maintenance I use monitoring tools so no need to waste time by regular manual monitoring.
I remember there was similar survey here a few years ago.
4 nodes on 1 PC.
2 active + 2 limited to 500GB that were added during test data period. I plan to GE them now they’re >6 months old.
Everyday: Look at StorjNodeViewer to confirm uptime & maybe look at ingress charts.
Everyday: Look at the forums for news.
Once a month: Log in to the VM they run on - typically after other maintenance/restart (yay Windows host) to make sure all is good.
Once a month: Check wallet
After the crazy days of test data, adding drives, rushing to migrate file systems to solve performance issues, and expecting a call from my ISP for bandwidth abuse any day, it’s all been smooth and quiet. Almost too quiet.
I operate just under 100 nodes, and spend way, way to much time on them
Around 80 of them are just 500GB max nodes, which are in various stages of filled. I enjoy the “challenge” of moving them about, testing different hardware with them, and observing the ingest patterns.
I have multiple locations each with multiple IPs, and run everything from large 500GB RAM servers to Raspberry Pis - the last of which is giving me nothing but trouble.
3 nodes setup with 10 TB each. No maintenance required, they’re all running in a Kubernetes environment with ceph backed storage. I probably check them quarterly for free space, but that’s about it. A while ago I had to move their backing storage from cephfs to RBD after the massive influx of tiny files during stress testing annoyed my MDSes.
They end up averaging ~93% download and ~99% upload success rates. It’s kind of sad actually. They’ve been up for 56, 30, and 26 months, and between the 3 there’s only ~10 TB used. There’s just so much churn, I get about 2TB per month of ingress, but I’ve been at 10TB since everything got cleaned up after stress testing.
I’ve got 5 nodes and 4 of them are problem free. 1 is still showing disqualified for some random thing that I cannot figure out. Uptime is nearing 100% but this Synology box is just weird.
