Read repair - configure to repair primary only #1844
Conversation
By default, the behaviour should be unchanged. however it is now configurable to read repair primary vnodes only - fallback vnodes will not be repaired on failing GETs, they will only receive new PUTs. See schema change for more details.
Initially to troubleshoot in test - but perhaps of generic use.
Rather than replicate piecemeal the standard PUT code within do_diffobj_put/3, instead use do_handoff_put/3 and use standard prepare_put/2 and perform_put/3 functions used in normal PUTs. The effect of this is that any optimisations in the normal PUT workflow, will now automatically be used for handoffs. Of particular relevance at this point is the HEAD (not GET) before PUT optimisation available with leveled backend. If there are large objects, and objects which already exist in the receiving vnode are to be handed off (such as in hinted handoff), then this increases efficiency. Some spec improvements to help with some editors that do not like fun() type. Some indent reductions to improve readability.
|
The change from This is (as per usual) a bucket type issue. The original assumption of bucket types is that they should be created and activated, only when the cluster is fully formed. However, we may need to expand (join nodes) at any time. As a node joins, the bucket types might not be propagated to the node before the handoffs start. In the case of the new The previous It is not clear that if PUTs immediately after joins could also have similar issues. This may not be exclusive to handoff. |
|
Need to consider UpdateHooks as well - https://github.com/basho/riak_kv/blob/riak_kv-3.0.12/src/riak_kv_vnode.erl#L3665-L3667 - if we use the general PUT code should the reason by a PutArg, and passed through? Only impacts the deprecated Yokozuna. |
This allows the same code to be used for both handoff and put.
As riak_core updated to ensure bucket types are exchanged prior to join committing
To make recovery of nodes easier, adding some helper functions to riak_client.
Remove legacy thumbs
* Read repair - configure to repair primary only
By default, the behaviour should be unchanged. however it is now configurable to read repair primary vnodes only - fallback vnodes will not be repaired on failing GETs, they will only receive new PUTs.
See schema change for more details.
* get_option returns value not {K, V}
* Add ability to suspend AAE
* Add logging of read repairs
Initially to troubleshoot in test - but perhaps of generic use.
* Handle handoff put through standard put code
Rather than replicate piecemeal the standard PUT code within do_diffobj_put/3, instead use do_handoff_put/3 and use standard prepare_put/2 and perform_put/3 functions used in normal PUTs.
The effect of this is that any optimisations in the normal PUT workflow, will now automatically be used for handoffs. Of particular relevance at this point is the HEAD (not GET) before PUT optimisation available with leveled backend. If there are large objects, and objects which already exist in the receiving vnode are to be handed off (such as in hinted handoff), then this increases efficiency.
Some spec improvements to help with some editors that do not like fun() type. Some indent reductions to improve readability.
* Make HookReason part of PutArgs
This allows the same code to be used for both handoff and put.
* Revert defaulting of properties
As riak_core updated to ensure bucket types are exchanged prior to join committing
* Add helpful operator functions to riak_client
To make recovery of nodes easier, adding some helper functions to riak_client.
* Update branches
Remove legacy thumbs
* Update rebar.config
By default, the behaviour should be unchanged. however it is now configurable to read repair primary vnodes only - fallback vnodes will not be repaired on failing GETs, they will only receive new PUTs.
See schema change for more details.
basho/riak_test#1369