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
Led to OOM kills, servers getting bounced. Was worse under libc malloc than jemalloc but it definitely still happens under jemalloc.
Tracked it down finally by very gently loading up a local instance of the server running in release mode + debug symbols w/ libc malloc under heaptrack with watch running curl at interval -n 1 in a few terminal tabs. Using drill made the server thrash and crash.
Seeing leaks that trace back to register_observer_handle_cleanup, ResourceData, get_state, etc. The common roots seem to be QueryState's .clone() and ResourceData's.
I'll be trying to figure out the least invasive way to disable leptos_query in the server so I can stop the bleeding.
The text was updated successfully, but these errors were encountered:
leptos-query/query/src/use_query.rs
Line 211 in b459583
Led to OOM kills, servers getting bounced. Was worse under libc malloc than jemalloc but it definitely still happens under jemalloc.
Tracked it down finally by very gently loading up a local instance of the server running in release mode + debug symbols w/ libc malloc under
heaptrack
withwatch
running curl at interval-n 1
in a few terminal tabs. Using drill made the server thrash and crash.Seeing leaks that trace back to
register_observer_handle_cleanup
,ResourceData
,get_state
, etc. The common roots seem to beQueryState
's.clone()
andResourceData
's.I'll be trying to figure out the least invasive way to disable
leptos_query
in the server so I can stop the bleeding.The text was updated successfully, but these errors were encountered: