Open tuxcanfly opened 8 months ago
Does it constantly crash or it was a single event?
Nvm, found the answer
The problem arises from the badger-db internal allocator, which failed to acquire memory. One possible reason for this issue could be that the store was initialized with a larger Memtable size than it actually ran on. This discrepancy might be due to the MemTable size being downscaled in the v0.12.1 release.
Is there a possibility that you upgraded the node application from a version pre-dating v0.12.1 without reinitializing the store?
Hi @walldiss I didn't remember which version I upgraded from, so I assume it was from v0.12.0. I tried running v0.12.0 with the same store, then upgraded to v0.12.2 and tried again. It didn't crash and I'm no longer able to reproduce this.
Also, is it generally recommended to reinit the store on minor version upgrades or major version upgrades or both? I'm asking because this could be useful information for service providers assuming not reiniting causes issues like this.
It also happens without upgrading the version.
But the last time I upgraded to v0.12.3 it worked. Possibly this behavior is random.
the same with me
Facing same issue for 0.12.4
Celestia Node version
v0.12.2
OS
macOS 14.1.0
Install tools
No response
Others
celestia-da
e1ad7cf4
Steps to reproduce it
it crashed after launch and keeps crashing, probably due to using the same node store
Expected result
node launches without crash
Actual result
node crashes on launch with the following log:
Relevant log output
Notes
No response