Thanks Ed, I could have done that but at this point my destination cluster is not healthy and I need to reinitialize that cluster, eventually the same tables will be there after initialization but with different table id. Once that happens I will setup replication again. At this point, i am not worried about in-flight replication data and I don't want this replication process to impact my primary cluster.. How safely can I achieve this?
-S ________________________________ From: dev1 <d...@etcoleman.com> Sent: Thursday, February 17, 2022 8:24 AM To: 'user@accumulo.apache.org' <user@accumulo.apache.org> Subject: [External] RE: how to stop entire (or on a single table) replication properly SO, I’m not familiar with replication in an operational setting – so my comments are based on my mental model of what I think replication is doing – the implement may not match my mental model – maybe someone else with more familiarity can chime in. I’m reading that you want to stop replication and do not care to preserve data that may be “in-flight” Why don’t you just stop replication on the source and then create the destination table that is expected to exist as the destination. When that data has been “replicated”, the source replication table should be empty – then just delete the destination table? You are still getting ride of the data and you let replication do the housekeeping for you? Ed Coleman From: Ligade, Shailesh [USA] <ligade_shail...@bah.com> Sent: Thursday, February 17, 2022 8:15 AM To: 'user@accumulo.apache.org' <user@accumulo.apache.org> Subject: Re: how to stop entire (or on a single table) replication properly Thanks Ed, Let me rephase it. I need to stop replication as my tables on the peer are changing. After stopping, I will need to start replication again to the tables. To stop the replication, on the primary instance tables i am going to set config to set replication false. Basically running config -t my_table -s table.replication=false (currently true). I believe that setting will stop replicating that table to peer. However, there is still data in primary replication table and system will still try to replicate to peer (on peer corresponding tables no longer exist!), i can see it is still replicating to the peer on the replication page on the monitor UI. I can set primary replication table offline, but when I bring it online again, that data will be still there. So the question is, how can I safely remove the data in primary replication table? One time i tried to do deleteall on primary replication table but when accumulo master re-started, it was complaining a lot about replication data, so just wanted to figure out proper steps. thanks -S ________________________________ From: dev1 <d...@etcoleman.com<mailto:d...@etcoleman.com>> Sent: Thursday, February 17, 2022 8:04 AM To: 'user@accumulo.apache.org' <user@accumulo.apache.org<mailto:user@accumulo.apache.org>> Subject: [External] RE: how to stop entire (or on a single table) replication properly I do not understand what you are asking – it would help if you stated what you are trying to accomplish and if you clearly identified source vs. destination. Ed Coleman From: Ligade, Shailesh [USA] <ligade_shail...@bah.com<mailto:ligade_shail...@bah.com>> Sent: Thursday, February 17, 2022 7:37 AM To: user@accumulo.apache.org<mailto:user@accumulo.apache.org> Subject: how to stop entire (or on a single table) replication properly Hello, If i must stop entire replication, I set config for an individual table replication to false. However this will not affect entries in the replication table and the system will keep (or try to keep) replicating. I can take replication table offline, but eventually when I need to start replication it will not clean. How can I delete entries in the replication table? Can i just do delete all? will that work? -S
