[infinispan-dev] CacheLoaders, Distribution mode and Interceptors

Mircea Markus mmarkus at redhat.com
Tue Mar 19 12:30:04 EDT 2013


On 19 Mar 2013, at 16:15, Dan Berindei wrote:

> Hi Sanne
> 
> On Tue, Mar 19, 2013 at 4:12 PM, Sanne Grinovero <sanne at infinispan.org> wrote:
> Mircea,
> what I was most looking forward was to you comment on the interceptor
> order generated for DIST+cachestores
>  - we don't think the ClusteredCacheLoader should be needed at all
> 
> Agree, ClusteredCacheLoader should not be necessary.
> 
> James, if you're still seeing problems with numOwners=1, could you create an issue in JIRA?
> 
>  
>  - each DIST node is loading from the CacheLoader (any) rather than
> loading from its peer nodes for non-owned entries (!!)
> 
> 
> Sometimes loading stuff from a local disk is faster than going remote, e.g. if you have numOwners=2 and both owners have to load the same entry from disk and send it to the originator twice. 
the staggering of remote gets should overcome that. 
> 
> Still, most of the time the entry is going to be in memory on the owner nodes, so the local load is slower (especially with a shared cache store, where loading is over the network as well).
+1
> 
>  
> This has come up on several threads now and I think it's critically
> wrong, as I commented previously this also introduces many
> inconsistencies - as far as I understand it.
> 
> 
> Is there a JIRA for this already?
> 
> Yes, loading a stale entry from the local cache store is definitely not a good thing, but we actually delete the non-owned entries after the initial state transfer. There may be some consistency issues if one uses a DIST_SYNC cache with a shared async cache store, but fully sync configurations should be fine.
> 
> OTOH, if the cache store is not shared, the chances of finding the entry in the local store on a non-owner are slim to none, so it doesn't make sense to do the lookup.
> 
> Implementation-wise, just changing the interceptor order is probably not enough. If the key doesn't exist in the cache, the CacheLoaderInterceptor will still try to load it from the cache store after the remote lookup, so we'll need a marker  in the invocation context to avoid the extra cache store load.
if the key does't map to the local node it should trigger a remote get to owners (or allow the dist interceptor to do just that)
> Actually, since this is just a performance issue, it could wait until we implement tombstones everywhere.
Hmm, not sure i see the correlation between this and tombstones? 

> 
> BTW your gist wouldn't work, the metadata cache needs to load certain
> elements too. But nice you spotted the need to potentially filter what
> "preload" means in the scope of each cache, as the metadata one should
> only preload metadata, while in the original configuration this data
> would indeed be duplicated.
> Opened: https://issues.jboss.org/browse/ISPN-2938
> 
> Sanne
> 
> On 19 March 2013 11:51, Mircea Markus <mmarkus at redhat.com> wrote:
> >
> > On 16 Mar 2013, at 01:19, Sanne Grinovero wrote:
> >
> >> Hi Adrian,
> >> let's forget about Lucene details and focus on DIST.
> >> With numOwners=1 and having two nodes the entries should be stored
> >> roughly 50% on each node, I see nothing wrong with that
> >> considering you don't need data failover in a read-only use case
> >> having all the index available in the shared CacheLoader.
> >>
> >> In such a scenario, and having both nodes preloaded all data, in case
> >> of a get() operation I would expect
> >> either:
> >> A) to be the owner, hence retrieve the value from local in-JVM reference
> >> B) to not be the owner, so to forward the request to the other node
> >> having roughly 50% chance per key to be in case A or B.
> >>
> >> But when hitting case B) it seems that instead of loading from the
> >> other node, it hits the CacheLoader to fetch the value.
> >>
> >> I already had asked James to verify with 4 nodes and numOwners=2, the
> >> result is the same so I suggested him to ask here;
> >> BTW I think numOwners=1 is perfectly valid and should work as with
> >> numOwners=1, the only reason I asked him to repeat
> >> the test is that we don't have much tests on the numOwners=1 case and
> >> I was assuming there might be some (wrong) assumptions
> >> affecting this.
> >>
> >> Note that this is not "just" a critical performance problem but I'm
> >> also suspecting it could provide inconsistent reads, in two classes of
> >> problems:
> >>
> >> # non-shared CacheStore with stale entries
> >> If for non-owned keys it will hit the local CacheStore first, where
> >> you might expect to not find anything, so to forward the request to
> >> the right node. What if this node has been the owner in the past? It
> >> might have an old entry locally stored, which would be returned
> >> instead of the correct value which is owned on a different node.
> >>
> >> # shared CacheStore using write-behind
> >> When using an async CacheStore by definition the content of the
> >> CacheStore is not trustworthy if you don't check on the owner first
> >> for entries in memory.
> >>
> >> Both seem critical to me, but the performance impact is really bad too.
> >>
> >> I hoped to make some more tests myself but couldn't look at this yet,
> >> any help from the core team would be appreciated.
> > I think you have a fair point and reads/writes to the data should be coordinated through its owners both for performance and (more importantly) correctness.
> > Mind creating a JIRA for this?
> >
> >>
> >> @Ray, thanks for mentioning the ClusterCacheLoader. Wasn't there
> >> someone else with a CacheLoader issue recently who had worked around
> >> the problem by using a ClusterCacheLoader ?
> >> Do you remember what the scenario was?
> >>
> >> Cheers,
> >> Sanne
> >>
> >> On 15 March 2013 15:44, Adrian Nistor <anistor at redhat.com> wrote:
> >>> Hi James,
> >>>
> >>> I'm not an expert on InfinispanDirectory but I've noticed in [1] that
> >>> the lucene-index cache is distributed with numOwners = 1. That means
> >>> each cache entry is owned by just one cluster node and there's nowhere
> >>> else to go in the cluster if the key is not available in local memory,
> >>> thus it needs fetching from the cache store. This can be solved with
> >>> numOwners > 1.
> >>> Please let me know if this solves your problem.
> >>>
> >>> Cheers!
> >>>
> >>> On 03/15/2013 05:03 PM, James Aley wrote:
> >>>> Hey all,
> >>>>
> >>>> <OT>
> >>>> Seeing as this is my first post, I wanted to just quickly thank you
> >>>> all for Infinispan. So far I'm really enjoying working with it - great
> >>>> product!
> >>>> </OT>
> >>>>
> >>>> I'm using the InfinispanDirectory for a Lucene project at the moment.
> >>>> We use Lucene directly to build a search product, which has high read
> >>>> requirements and likely very large indexes. I'm hoping to make use of
> >>>> a distribution mode cache to keep the whole index in memory across a
> >>>> cluster of machines (the index will be too big for one server).
> >>>>
> >>>> The problem I'm having is that after loading a filesystem-based Lucene
> >>>> directory into InfinispanDirectory via LuceneCacheLoader, no nodes are
> >>>> retrieving data from the cluster - they instead look up keys in their
> >>>> local CacheLoaders, which involves lots of disk I/O and is very slow.
> >>>> I was hoping to just use the CacheLoader to initialize the caches, but
> >>>> from there on read only from RAM (and network, of course). Is this
> >>>> supported? Maybe I've misunderstood the purpose of the CacheLoader?
> >>>>
> >>>> To explain my observations in a little more detail:
> >>>> * I start a cluster of two servers, using [1] as the cache config.
> >>>> Both have a local copy of the Lucene index that will be loaded into
> >>>> the InfinispanDirectory via the loader. This is a test configuration,
> >>>> where I've set numOwners=1 so that I only need two servers for
> >>>> distribution to happen.
> >>>> * Upon startup, things look good. I see the memory usage of the JVM
> >>>> reflect a pretty near 50/50 split of the data across both servers.
> >>>> Logging indicates both servers are in the cluster view, all seems
> >>>> fine.
> >>>> * When I send a search query to either one of the nodes, I notice the following:
> >>>>   - iotop shows huge (~100MB/s) disk I/O on that node alone from the
> >>>> JVM process.
> >>>>   - no change in network activity between nodes (~300b/s, same as when idle)
> >>>>   - memory usage on the node running the query increases dramatically,
> >>>> and stays higher even after the query is finished.
> >>>>
> >>>> So it seemed to me like each node was favouring use of the CacheLoader
> >>>> to retrieve keys that are not in memory, instead of using the cluster.
> >>>> Does that seem reasonable? Is this the expected behaviour?
> >>>>
> >>>> I started to investigate this by turning on trace logging, in this
> >>>> made me think perhaps the cause was that the CacheLoader's interceptor
> >>>> is higher priority in the chain than the the distribution interceptor?
> >>>> I'm not at all familiar with the design in any level of detail - just
> >>>> what I picked up in the last 24 hours from browsing the code, so I
> >>>> could easily be way off. I've attached the log snippets I thought
> >>>> relevant in [2].
> >>>>
> >>>> Any advice offered much appreciated.
> >>>> Thanks!
> >>>>
> >>>> James.
> >>>>
> >>>>
> >>>> [1] https://www.refheap.com/paste/12531
> >>>> [2] https://www.refheap.com/paste/12543
> >>>> _______________________________________________
> >>>> infinispan-dev mailing list
> >>>> infinispan-dev at lists.jboss.org
> >>>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
> >>>
> >>> _______________________________________________
> >>> infinispan-dev mailing list
> >>> infinispan-dev at lists.jboss.org
> >>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
> >> _______________________________________________
> >> infinispan-dev mailing list
> >> infinispan-dev at lists.jboss.org
> >> https://lists.jboss.org/mailman/listinfo/infinispan-dev
> >
> > Cheers,
> > --
> > Mircea Markus
> > Infinispan lead (www.infinispan.org)
> >
> >
> >
> >
> >
> > _______________________________________________
> > infinispan-dev mailing list
> > infinispan-dev at lists.jboss.org
> > https://lists.jboss.org/mailman/listinfo/infinispan-dev
> _______________________________________________
> infinispan-dev mailing list
> infinispan-dev at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/infinispan-dev
> 
> _______________________________________________
> infinispan-dev mailing list
> infinispan-dev at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/infinispan-dev

Cheers,
-- 
Mircea Markus
Infinispan lead (www.infinispan.org)







More information about the infinispan-dev mailing list