Updates on Test Data

[pangolin]

@Andrew

forgot to roll out this version to the nodes? Really?

We just witnessed a major bug been rolled out to slc sat. If this was rolled out to production sats it would have been a catastrophe. So let them do it cautiously.

[MarviBiene]

@Alexey

I’m sorry, but for that you need to be a customer. You are a supplier, I’m sorry!

So if I understand it right, we as supplier aren’t eligible for reliable and right data?

[littleskunk]

@MarviBiene

So if I understand it right, we as supplier aren’t eligible for reliable and right data?

Switch over to grafana is the solution.

[MarviBiene]

I tried it, but as mentioned in another post my Prometheus instances are killing my computer. The ram and disk usage are exploding after two days and I don’t know how to fix this

[littleskunk]

Increase interval and shorten the history will help.

[ACarneiro]

@pangolin

I can see GC deleting massive amounts of 1-2 weeks old data from slc. This data should be 30 days TTL so how is this possible?

Same here. One of my nodes has put almost half the uploaded data into trash. I am following this with great interest…

[Mitsos]

If TTL data is moved to trash before it expires, that means it was deleted (=client hitting delete). It is then considered trash and moved to trash by a GC. Nothing changes on how data is manually deleted because it is no longer needed, even if it is TTL data. TTL data just has an expiry date, that’s all.

[daki82]

Today i have been told by my wife, that our iptv was lagging since weeks.
stopping the node fixed it.

Did a priorisation in the router for the Reciver, diddn’t help. set max conkurrent to 5.
testing ongoing.

v1.104.5 wouldn’t stop at first, reboot of the pc needed.

[peter_linder]

You will need to explain to your wife the importance of node operation. Tell her to calm down if that doesn’t help

[arrogantrabbit]

To keep beating the dead horse — have you tied SQM? Prioritization won’t help if you have bufferbloat

[ACarneiro]

@daki82

v1.104.5 wouldn’t stop at first, reboot of the pc needed.

This is interesting. How long did you wait for the process to stop?
Is this Windows or Linux? Docker container?

[Roxor]

@peter_linder

You will need to explain to your wife the importance of node operation. Tell her to calm down if that doesn’t help

…and then when you’re a bachelor again… living on your own… you’ll be able to run as many nodes as you want!

[ACarneiro]

@Roxor

you’ll be able to run as many nodes as you want!

And you’ll have nothing but the heat of those nodes to keep you warm on those lonely nights…

[Aitor]

You tell us that you need more nodes but it seems that with the current ones there are many things that fail. Today my trash has increased by 1,23 TB.

[daki82]

windows native, service window, green line walked all the way, maybe one minute.
then it was all options grey in the service panel. task manager process still visible, also drive working 100%, stopped primocache to be sure, did not help.

@arrogantrabbit

have you tied SQM

i googled it,(expecting corn) and it made no sense at all
if storj bloating my 7590AX i would leave the project, selling the storj-hardware.

@Roxor

@peter_linder

You will need to explain to your wife the importance of node operation. Tell her to calm down if that doesn’t help

…and then when you’re a bachelor again… living on your own… you’ll be able to run as many nodes as you want!

both to good to be true, sadly the 2 i have is already the limit. until i make 40 or 60 per month for the second internet line.

max concurrent 5 worked so far. cpu load dropped from 9 to under 1 %
going to mcr 10 now, cpu doubled also. interesting.

My piece expiration db is now 6GB big… adjusted primocache

[arrogantrabbit]

@daki82

if storj bloating my 7590AX i would leave the project, selling the storj-hardware.

lol tried

It’s not your gateway, it’s what’s upstream of your gateway. Modem, or ISP last mile equipment etc. On the contrary, your gateway does indeed need to be quite beefy to run the SQM algorithms at reasonable speeds. But it’s worth it. I have gigabit downstream, docsis3 modem, that already has some qos, and still, SmartQueues on ubiquiti help further reduce already non -unreasonable latency under load.

[Toyoo]

@Roxor

@MarviBiene

And we as node operator are in my opinion the same as paying customers.

SNOs are service providers: we don’t pay Storj like customers: we get paid. And payouts have been correct month-after-month. If we don’t feel we’re getting paid adequately: we stop providing our service (and find another project)

@Alexey

I’m sorry, but for that you need to be a customer. You are a supplier, I’m sorry!

@MarviBiene’s argument is wrong, but the conclusion is actually right. As suppliers we should be considered business partners. If our business partner lies to us, then this is an increased risk to us as suppliers. Numbers that the dashboards show is communication from Storj to SNOs. If they are systematically incorrect, this is wrong.

@littleskunk

I am not 100% sure yet but it looks like the uploads didn’t had enough randomness in the filenames and some uploads got overwritten.

Happy to hear you have found the issue. I would love to know what was the root cause (just coding in haste, or a wrong decision somewhere?), this sounds like a nice case study.

@Alexey

Do you want to help? We are hiring: Careers or you may submit a PR, our devs are happy to accept the Community contribution!

I only see an SRE position open. Do you want to say you also look for software engineers? I would apply.

[Roxor]

@Toyoo

As suppliers we should be considered business partners. If our business partner lies to us, then this is an increased risk to us as suppliers. Numbers that the dashboards show is communication from Storj to SNOs. If they are systematically incorrect, this is wrong.

We’re not “business partners”. Storj is our customer. We agree to run their software and provide space/connectivity, and they pay us. We’re in complete control of our environments: can measure any data going in or out of it, and any space used within. If Storj offers us some optional reports in the Node UI, then we can choose to use them at our convenience. Or report on things ourselves with our own tools.

If we’re unhappy with Storj software, or feel we aren’t being paid accurately, we can drop them as a customer. Storj knows us only as Node Identities, random email addresses, and opaque ETH wallets - that’s nowhere near a “business partner” relationship. We’re strangers exchanging services for coins.

And I’m fine with that. If they pay me every month, then they can remain my customer

[Toyoo]

@Roxor

Storj knows us only as Node Identities, random email addresses, and opaque ETH wallets - that’s nowhere near a “business partner” relationship.

That is not what this page suggests. If they are our customers, why do they demand tax forms from us?

[Roxor]

That’s the US government docs for reporting income earned in the US. SNOs are supposed to fill them out on their own… but nobody does… so Storj requests them to cover-their-butts (if audited). It’s for taxes owed (or not owed) to the US government. They are not business partner paperwork for Storj.