Title: Windows node stops after manual v1.155.8 update — corrupted/unreadable folder inside trash
Hi everyone,
I am running a Storj storage node on Windows, installed directly on drive E:\.
After the recent updater issue with v1.155.8, I followed the workaround mentioned in the forum:
-
Downloaded storagenode_windows_amd64.zip from v1.155.8
-
Downloaded storagenode-updater_windows_amd64.zip from v1.155.4
-
Stopped the Storj services
-
Backed up the old executables
-
Replaced only:
E:\storagenode.exe
E:\storagenode-updater.exe
The service configuration currently points to:
storagenode:
E:\storagenode.exe run --config-dir “E:\\”
storagenode-updater:
E:\storagenode-updater.exe run --config-dir “E:\\” --binary-location “E:\storagenode.exe” --log.output “winfile:///E:\storagenode-updater.log”
However, the storagenode service starts and then stops automatically.
When I run the node manually from PowerShell:
cd E:\
.\storagenode.exe run --config-dir “E:\\” --log.output stdout
the node initially starts successfully. It loads the configuration, opens the hashstore, starts the public server on port 28967, and begins uploading/downloading pieces. Then it shuts down with this error:
ERROR piecestore:cache error getting current used space for trash:
filestore error: failed to walk trash namespace …
GetFileInformationByHandleEx E:\trash\qstuylguhrn2ozjv4h2c6xpxykd622gtgurhql2k7k75wqaaaaaa\2026-05-31\cu:
The file or directory is corrupted and unreadable.
ERROR services unexpected shutdown of a runner
name: piecestore:cache
ERROR failure during run
filestore error: failed to walk trash namespace …
E:\trash\qstuylguhrn2ozjv4h2c6xpxykd622gtgurhql2k7k75wqaaaaaa\2026-05-31\cu:
The file or directory is corrupted and unreadable.
So it looks like the manual update worked, but now the node cannot stay running because one folder/file inside E:\trash appears to be corrupted or unreadable.
My questions are:
- Is it safe to delete only this specific old trash folder?
E:\trash\qstuylguhrn2ozjv4h2c6xpxykd622gtgurhql2k7k75wqaaaaaa\2026-05-31
-
Should I run chkdsk E: /f first before deleting anything?
-
Is there any recommended Storj procedure for this situation on Windows, to avoid damaging the node data or reputation?
I have not touched blobs, orders, hashstore, config.yaml, or any identity/data folders. I only replaced the two .exe files mentioned above.
Any guidance from the developers or experienced SNOs would be appreciated. Thanks.