Hello Everyone, a week ago i got several errors on my main node. Maybe someone can help me to solve the problem. The Node runs on Windows 10. Thanks for your Help. 2023-03-27T05:11:25.897+0200 INFO piecestore download started {Piece ID: 7WMQ364BMF672RLAQWDD76IQ6EORNXMYXIL5PGN46I4VUT4P4JXA, S…

[image] Craig: In 1.75.2 a change was made to report disk timeouts at a period that is either shorter than before or had not been set before at all. In turn it is throwing errors on nodes that previously indicated no problem because they weren’t being “caught” for being slow. [im…

it’s a windows 10 machine - where/which log should I check? And yea, there’s still enough free space on it: [image]

I’d open Event Viewer, and look in the Custom Views → Administrative Event, and Windows Logs → System logs. Those logs are often full of crap that you don’t really have to worry about: but maybe the System one will show errors related to your disk?

Honestly I didn’t see anything there that would indicate the disk being involved…

Amount of disk IO in the resource monitor? Disk response latency when this happens? Looks like your disk can’t keep up when IO slightly increases. Look into PrimoCache. Also run check disk to verify filesystem integrity.

which flags should I run for the checkdisk? It’s a standalone drive just for the node, not the filesystem. The IO load seems to be fine? [image]

Its related to slow disk subsystem or fragmentation. Did you ever defrag the node? Also various solutions here: [image] Fatal Error on my Node troubleshooting Hello Everyone, a week ago i got several errors on my main node. Maybe someone can help me to solve the problem…

[image] The IO load seems to be fine? Looks fine right now. But you are not seeing issues right now, right? You may want to monitor IOPs and latency on the disk long term (e.g. with performance monitor) and when the issue reproduces review what was the load at the time issue has occurred. Be…

[image] arrogantrabbit: This produces a very small improvement if any, and is not worth wasting time doing. Any proof for that? And why do some people here recommend it? NTFS with nodes mess this up over time, pushing seek time up, when walking the millions of files.

I am defragging a different drive atm, but they are completely independent. And my cpu usage is below 25%

The databases are on a different SSD, so the drives only deal with the data itself. I’ll run the check disk command now though

Fatal Error on my Node / Timeout after 1 min

Node Operators troubleshooting

daki82 December 19, 2023, 7:22pm 504

Its related to slow disk subsystem or fragmentation.

Did you ever defrag the node?
Also various solutions here:

Locks like you defrag the D drive atm.