since some STORJ-Versions the Windows Node keeps crashing. It runs for like a day, then crashes and I need to restart it so it lasts another day. I checked the logs and they looked fine with all the INFO messages with sucessful up- and downloads.
Can you tell what may be the reason for the constant crashes?
I guess it’s likely the same issue as discussed in
And solution would be:
Stop the node
Check the disk for errors and fix them
Run defragmentation, make sure that the automated defragmentation is enabled for this drive
Check for timeouts after a while
If the node would stop with a timeout again - increase the related timeout and restart
If the timeout would be too great (more or equal 5 minutes), it’s time to check what’s wrong with that drive or your setup, perhaps it has hidden issues, which should be addressed.
It is this error message and the node crashed again:
FATAL Unrecoverable error {“error”: “piecestore monitor: timed out after 1m0s while verifying writability of storage directory”, “errorVerbose”: “piecestore monitor: timed out after 1m0s while verifying writability of storage directory\n\tstorj.io/storj/storagenode/monitor.(*Service).Run.func2.1:169\n\tstorj.io/common/sync2.(*Cycle).Run:160\n\tstorj.io/storj/storagenode/monitor.(*Service).Run.func2:161\n\tgolang.org/x/sync/errgroup.(*Group).Go.func1:75”}
Am now checking the disk with the command:
chkdsk E: /F
It must be written in capital letters otherwise it can’t find it.
Edit: How often should the defragmentation run? Once in a month is enough?
Then perhaps your node collected enough data to slow down your disk (if it doesn’t have issues though), so you may try to increase a timeout for a writeability check in your case.
on healthy node no problem.
on slow node you can set it to full for less disturbing write while defrag.
atm my node with databases on ssd fragments at 0.2% per 50GB disk space used.or al 2days
so all 40-60 days defrag is ok i think. maybe 3 times a year also.
on my 7.8TB node with databases on disk. defrag will run for one whole week.
This is a bad practice, because the regular defragmentation will be shorter than if you run it once in several months especially on the storage location, where pieces moves very often (if it still has a free space).
right, it has a low priority.
on SSD the fragmentation has almost no impact on latency, since there is no mechanical moves.
I strongly disagree. Unlike ext4 the NTFS fs has a serious impact on latency due fragmentation.
So, please do not disable it for the storage location.
To support my opinion I can invite @Vadim to confirm that defragmentation has fixed an issue with disk timeouts for his setup.