earthcube / geocodes

This is the containers stacks to run geocodes
0 stars 0 forks source link

blazegraph truncate journal #33

Closed valentinedwv closed 10 months ago

valentinedwv commented 1 year ago

The famous, hey, can we document the exact steps for our systems, and not just the see here, so we can do this repeatedly, with a script request to truncate the blazegraph journal.

total 5823840
-rw-r--r--    1 root     root     6583746560 Jul 17 18:31 blazegraph.jnl
-rw-r--r--    1 root     root             0 May 15 18:52 rules.log
valentinedwv commented 1 year ago

MB managed this a one point.

Question is can this be documented on how to do it on our present stacks,

so we can have a repeatable procedure, and maybe even a shell script.

ylyangtw commented 1 year ago

also check blazegraph truncate journal's usage

valentinedwv commented 1 year ago

Thought on possibly scripting this using the docker/portainer api... once we know if the start/stop technique works

using docker.py module

valentinedwv commented 1 year ago

I think i was reading this wrong, and based on experience with Fuseki issues with journal filling disk.

Reading the doc says it will recover journal when it hits 200m statements.

I might be misunderstanding how the journaling works... might just be we have not hit the recylce limits. https://github.com/blazegraph/database/wiki/CommonProblems#problem-i-am-using-the-journal-and-the-file-size-grows-very-quickly 12:58 Looks like we are using the rw store this is from the docker file. https://github.com/nawerprod/blazegraph/blob/master/RWStore.properties

ylyangtw commented 1 year ago

putting this off. we probably don't need to worry about this.