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
we have some events there, e.g.:
the event shows that there was a disruption: disrupt_remove_node_then_add_node at the time of the event (2024-09-10 16:59:35.814)
but from Nemesis tab we can learn that we had multiple disruptions running during this time:
both disrupt_remove_node_then_add_node and disrupt_create_index started before 2024-09-10 16:59:35.814 and finished after that.
So the event should show both disruptions. I think the implementation assumes there's at most one disruption at any given moment, but SCT has parallel nemeses.
look inside the event, during_nemesis= actuall show exactly what you are looking
yes the graphical one Argus showing isn't show everything, and should be taking it first from the event information itself (data created by SCT itself), and only then fallback to lookups in the nemesis information.
Some events happen during multiple nemeses. This commit updates the
logic to find all nemeses that were active during a specific event,
instead of just looking for one like previously.
Fixesscylladb#470
Some events happen during multiple nemeses. This commit updates the
logic to find all nemeses that were active during a specific event,
instead of just looking for one like previously.
Fixesscylladb#470
For example let's take https://argus.scylladb.com/test/5cd4bfe3-231b-4f48-99bc-f21ebd1b5efe/runs?additionalRuns%5B%5D=e330a762-61b1-4318-a55c-0b072d4df26e
we have some events there, e.g.:
the event shows that there was a disruption:
disrupt_remove_node_then_add_node
at the time of the event (2024-09-10 16:59:35.814
)but from
Nemesis
tab we can learn that we had multiple disruptions running during this time:both
disrupt_remove_node_then_add_node
anddisrupt_create_index
started before2024-09-10 16:59:35.814
and finished after that.So the event should show both disruptions. I think the implementation assumes there's at most one disruption at any given moment, but SCT has parallel nemeses.
cc @temichus
The text was updated successfully, but these errors were encountered: