This page is not created by, affiliated with, or supported by Slack Technologies, Inc.
I have my local tx log and I think about migrating it to some kind of persistent store, like postgres. Does this github issue mean that I can't do it easily? https://github.com/xtdb/xtdb/issues/1386 I don't talk about checkpoints because they're on-demand. I'm thinking about migrating the tx log. Or do you think that I shouldn't bother and instead create a snapshot?
How does checkpointing work if I want to have Lucene index but I'd enable this index already after the database has advanced a little bit? Would it try to reindex the unindexed documents or would it only index items from newly ingested tx log? i.e. would it pick up old docs?
But what if I would destroy the txlog and only then enable lucene? Then there would not be any tx log to pick it up :thinking_face:
https://docs.xtdb.com/administration/configuring/ as per the important section
> destroy the txlog? you would intentionally lose your data This doesn't answer the question. If I have a long tx log and make a snapshot, can I throw away the first part of it? If I don't lose the snapshot. i.e. I'd only have the snapshot and tx-log after it. I'm not trying to get rid of the tx log.
So even if I create a snapshot I should still have the tx log? Oh
yes… snapshots can be thrown away as they can always be recreated from the golden stores… but not the other way around
What I thought about was that I'd throw away the tx log before the snapshot and then the new index would have to somehow iterate through each indexed doc in the memory. It could also work but I was trying to make sure if this is the right idea.
I'll have another question about txlog storage in separate comment.