How do I check that everything is working well?

Node Operators FAQ
1

Hi!

Unfortunately, I have an SMR disk.
I decided to use it for storj since it’s useful for now.

I know that SMR disks may have very poor performance.
How can I check that the performance are good enough for Storj (I/O, etc.)? Do I have to wait for, maybe, being DQ? Or is there a better way to know it (via Storj logs for example but I don’t know what I would look for)?

Thanks a lot!

2

The IO wait is too much or you might see 100% cpu usage or database is locked message in the logs. If you can avoid using this disk you should else its a gamble. Although there are SNOs that are running a node smoothly with SMR drive so you might not even encounter any problem at all.

2 Likes
3

You likely will not be disqualified, but performance might be bad. This could cause database locked issues, which are more likely to lead to suspension than disqualification. It could also impact performance of anything else running on that system.

Not everyone is seeing problems with their SMR drives. I guess they’re not all made equal. We’ve also seen some indication that running multiple nodes on different HDD’s can remedy the issues with SMR by spreading the load across multiple HDD’s. So rather than replacing the HDD, I would suggest simply adding one as a new node, if that’s an option.

1 Like
4

Thanks.
What’s the impact of a “suspended” node?

5

Suspension acts a bit like temporary disqualification. Your node will not receive uploads while in suspension, but if you manage to fix the problems causing the suspension it can resume normal operation. This happens when your node is having audit issues that aren’t related to missing data or data corruption.

After successfully responding to a few audits the suspension mode will disappear and the node will get uploads again.

6

You dont have any interection with the sattelite until the satellite decide your node is ok again. In my experience it can take hours. I also noticed that the suspended message for a given satelite is not accurate because when i look logs tail in real time i can see interaction with a so called suspended one.

Also the devs have flaged it as a known issue and are working on it :

1 Like
7

Thanks for this explanation.
If don’t receive email about suspension and if I dont have any notification about it in my dashboard, it should be fine, right?

8

Yes, you would be notified if this is an issue. With SMR drives the interesting part is that suspension would stop uploads, which would lower the load on the HDD, give it time to write CMR cache to SMR tracks and it would likely recover automatically. You don’t really want to flip in and out of suspension though, but it’s unlikely you get disqualified as a result of this.

1 Like