Couldn't another reason for doing cleanup sequentially be to avoid data loss? If data is being streamed from a node during bootstrap and cleanup is run too soon, couldn't you wind up in a situation with data loss if the new node being bootstrapped goes down (permanently)?
On Thu, Nov 28, 2013 at 8:59 PM, Aaron Morton <aa...@thelastpickle.com>wrote: > I hope I get this right :) > > Thanks for contributing :) > > a repair will trigger a mayor compaction on your node which will take up a > lot of CPU and IO performance. It needs to do this to build up the data > structure that is used for the repair. After the compaction this is > streamed to the different nodes in order to repair them. > > It does not trigger a major compaction, that’s what we call running > compaction on the command line and compacting all SSTables into one big > one. > > it will flush all the data to disk that will create some additional > compaction. > > The major concern is that s a disk IO intensive operation, it reads all > the data and writes data to new SSTables (a one to one mapping). If you > have all nodes doing this at the same time there may be some degraded > performance. And as it’s all nodes it’s not possible for the Dynamic Snitch > to avoid nodes if they are overloaded. > > Cleanup is less intensive than repair, but it’s still a good idea to > stagger it. If you need to run it on all machines (or you have very > powerful machines) it’s probably going to be OK. > > Hope that helps. > > ----------------- > Aaron Morton > New Zealand > @aaronmorton > > Co-Founder & Principal Consultant > Apache Cassandra Consulting > http://www.thelastpickle.com > > On 26/11/2013, at 5:14 am, Artur Kronenberg < > artur.kronenb...@openmarket.com> wrote: > > Hi Julien, > > I hope I get this right :) > > a repair will trigger a mayor compaction on your node which will take up a > lot of CPU and IO performance. It needs to do this to build up the data > structure that is used for the repair. After the compaction this is > streamed to the different nodes in order to repair them. > > If you trigger this on every node simultaneously you basically take the > performance away from your cluster. I would expect cassandra still to > function, just way slower then before. Triggering it node after node will > leave your cluster with more resources to handle incoming requests. > > > Cheers, > > Artur > On 25/11/13 15:12, Julien Campan wrote: > > Hi, > > I'm working with Cassandra 1.2.2 and I have a question about nodetool > cleanup. > In the documentation , it's writted " Wait for cleanup to complete on > one node before doing the next" > > I would like to know, why we can't perform a lot of cleanup in a same > time ? > > > Thanks > > > > > -- - John
