At start up time, that doesn’t surprise me as the file walker is running then. What’s weird is that the load remains consistent. You could try disabling the file walker and see if that changes anything.
Have you noticed your node restarting by itself or maybe low uptime numbers that suggest it has restarted? This could cause the file walker to keep triggering and might signal additional issues.
2 Likes
it is 5400 drive, YES it working litle slower. But it last longer, make less heat.
1 Like
I do have low uptime at this moment. - but that is because i have done changes to hardware (case and stuff) so i have beeen doinmg alot of node restarts. also windows fucked me and updated anyways so that is 100% also some restarts- but i think i have fully dissabled that now.
Can you guide my through how i dissable filewalker? or do you have a link for a guide?
Don’t you just love Windows…? And unless your using an enterprise version of windows, unless something has changed, I can almost guarantee you have not disabled windows updates even if you did… and windows loves to update and reboot CONSTANTLY!
1 Like
The relevant topic was linked here
This one is simple, you just add: storage2.piece-scan-on-startup: false to your config.yaml and then restart your node.
The file walker on restart has the primary function to see how much data is currently used by your node. Disabling it can make that numbers start going out of sync. If this step doesn’t help, I recommend enabling it again. When your HDD starts filling up it may be a good idea to run file walker again to make sure you don’t get unintended overusage of HDD space which could cause problems.
1 Like
I am so sick of it! and i cant switch to any other OS for other reasons -_- windows is so shit.
1 Like
Thanks for the answer - i will give it a go!
Be aware of that Toshiba doesn’t honour the warranty to end costumers. That reply I got from Toshiba back in 2020
Dear Donald Motsinger.
Sorry, we cannot accept your account request for ‘internal’ storage products!
These account requests are only intended for Toshiba authorized distributors.
Please contact your dealer for warranty processing.
So you’re at the mercy of the seller you bought it from. Ebuyer UK only gave me a partial refund for that drive. And that was it with Toshiba for me.
1 Like
Thanks alot for the headsup!
Unfortunately, many companies try this, but it is in many places illegal. They HAVE to offer manufacturer warranty for consumers as well. It’s not a good sign that they have this as standard business practice, but if you press them, they should cave.
You could always send a letter to the manufacturer by certified mail requesting a return receipt so you know they received it. Your letter should demand the repairs promised in the warranty. This legal paper trail is very likely to change their minds.
1 Like
Write cache is usually not recommended as it can corrupt the databases on an unexpected shutdown. However, if you move your DB’s to an SSD, the biggest risk with enabling write caching on the HDD is losing or corrupting maybe a piece or 2. But in most cases, that simply means the transfer wasn’t finished completely, so you’re node is not expected to keep that piece anyway.
If you click through the warranty processor here Warranty - EMEA Region – Toshiba Storage Solutions you’ll end up on this page StorRepair GmbH Service Portal
There you can see they only process external hard disks from end customers. It looks they only do this for the EMEA region. It seems to be different in the US
https://myapps.taec.toshiba.com/postSales/toshiba_welcome_page.jsp
I just vote with my wallet in the future before going through that hassle again…
Definitely not legal here… But yeah, you’re right. Vote with your wallet with these policies. It’s not acceptable.
@BrightSilence Now the node (3.3TB filled ish) has been on for over 24 hours and still pegged at 100% disk usage:
I dont understand why its running 100% with such slow reads and writes speeds… Its annoying. Is there somehow i can check in logs whats the disk is actually doing?
filewalker should not even take over 24 hours?
The read/write speeds are low and disk IO is high because of all the small files. The disk is pegged at 100% usage in terms of IO operations, not bandwidth. I’m currently copying a 1.5 TB node from a 4 TB disk to an 18 TB disk while the node is running. It’s currently at 1.35 TB copied after 38 hours and the current speed is about 11 MB/s. The speed varies though depending on file sizes, location on the disk, node activity etc. So your numbers are pretty normal there, and I wouldn’t be surprised if filewalker took over 24 hours on a 5400 RPM drive. We’re also getting some increased ingress lately so this will account for some extra disk IO as well.
Just keep an eye on your online / audit scores. As long as they’re not dropping you should be fine just waiting it out. If they do start dropping though it could be an indicator that the disk just can’t keep up. At that point I would disable the filewalker temporarily in order to keep it online and source a faster disk to copy the node to. Also keep in mind this will get worse the larger the node gets. More data means more to scan, but also as spinners fill up they get slower the further the data is to the outside of the platters.
Okay thanks… but if it is allready 100% running for soo long after just a single little node reboot. then i whould think it will be even worse on a 10TB full node…? is it even going to be a working node at 10TB - let even yours on 18TB? 
My largest node is about 13 TB and it still works fine. Not sure how long filewalker takes for that one though. A lot of people are concerned about 100% disk usage as well though. There’s been requests for the ability to tune it or lower the intensity etc. I think that would be a good idea but well have to wait and see if Storj impliments something like that. I would imagine having so many disks pegged at 100% can’t be good for the network especially in the long term.
1 Like
it really does not makes sense - disks also dies faster - creates mores heat - uses more power. makes it more expensive to be a SNO…
also - i suspect that even when filewalker is complete on mine - its still at 100%. but i will wait and see.
why cant we have some sort of “progress” on file walker - to see how far it is.
Did you end up moving the db’s to another drive? If so, you could try enabling write caching on the HDD. But don’t do this if the db’s are still on there, otherwise you risk db corruption.
There are some risks to this. Pieces could get corrupted on unsafe shutdowns. So if you have unreliable power or an unstable system, don’t do this.
2 Likes