You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Parity by default performs periodic snapshot activity once every 5000 blocks to verify
it's correctness and for other useful bits.
However when this happens, the node gets unresponsive, might stop responding to RPC calls and in general can make the node stop completely working.
The process doesn't exit, but log messages can make us see that the process is stuck syncing a specific snapshot without success. This also causes huge spikes in memory and disk access.
It is possible to sync (although a bit more slowly) without using snapshots.
My suggestion is to keep periodic snapshots until the node is synced once and then remove this flag from the way we run Parity programatically from some sub process that we run to audit it.
Or we can just run without the snapshots all together as well.
Parity by default performs periodic snapshot activity once every 5000 blocks to verify
it's correctness and for other useful bits.
However when this happens, the node gets unresponsive, might stop responding to RPC calls and in general can make the node stop completely working.
The process doesn't exit, but log messages can make us see that the process is stuck syncing a specific snapshot without success. This also causes huge spikes in memory and disk access.
It is possible to sync (although a bit more slowly) without using snapshots.
My suggestion is to keep periodic snapshots until the node is synced once and then remove this flag from the way we run Parity programatically from some sub process that we run to audit it.
Or we can just run without the snapshots all together as well.
Your thoughts well appreciated @erankirsh @OdedWx
The text was updated successfully, but these errors were encountered: