The loop to expire messages in filestore.goexpireMsgs() seems to take very long!
In part this has to do with updating of consumer meta data, that is made even slower by the wildcard filter_subject
Side note: the signal handler for ctrl-c seems to be set up after executing this code. Maybe set up before, so this can be stopped without killing.
This screenshot shows the loop with tracing modifications, so the output below makes sense
This output is showing roughly 20K messages expired every two minutes.
This is two traces on line 2196 of 10K each (line 2195) between unrelated traces that are 2 minutes apart.
For context this is seqno ~3M with a last seqno of ~9M.
consumer in the stream that is recovered contains a
filter_subject
for a wildcard subjectsingress.raw.*
The loop to expire messages in
filestore.go
expireMsgs()
seems to take very long! In part this has to do with updating of consumer meta data, that is made even slower by the wildcardfilter_subject
This screenshot shows the loop with tracing modifications, so the output below makes sense
This output is showing roughly 20K messages expired every two minutes. This is two traces on line 2196 of 10K each (line 2195) between unrelated traces that are 2 minutes apart. For context this is seqno ~3M with a last seqno of ~9M.