Setting both offHeapMaxMemory and cacheStoreFactory
I have a few doubts on off heap cache configuration 1. Is it possible to set both off-heap and cache store with write behind mode for a cache? The configuration looks like: 2. from which place (onheap or offheap or both) the cached objects are written to cache store (MySQL above) if write behind is enabled? 3. What will happen to a new item to be placed in cache when the size (both onheap and offheap) has already exceeded the limit? -- View this message in context: http://apache-ignite-users.70518.x6.nabble.com/Setting-both-offHeapMaxMemory-and-cacheStoreFactory-tp2527.html Sent from the Apache Ignite Users mailing list archive at Nabble.com.
Re: Serialization issue with Ignite 1.5 built from master (357d791)
> > Alexey, > > Do you mean this? > > class MyWrapper { >// Wrapped map. >Map<> myMap = …; > } > Yes, this is exactly what I meant.
putting to cache as generic object
In our project, we plan to deploy ignite caching as a shared service available for other modules. The caching module is visible to all the modules but any module's classes are not visible to caching. The objects of different components are put into cache, but it requires classes of different modules to be on Ignite's classpath.I plan to use Reflection like mechanism at ignite server to fetch the property of the cached object. But before this Ignite OptmizedMarshaller expects class to be present in classpath while serializing. For some security reason I can't put other modules classes into Ignite classpath and also can't use peerClassLoading. I want a solution which can make ignite to treat objects to be cached as generic 'java.lang.Object's and don't expect cached object's class to be present in Ignite's classpath -- View this message in context: http://apache-ignite-users.70518.x6.nabble.com/putting-to-cache-as-generic-object-tp2525.html Sent from the Apache Ignite Users mailing list archive at Nabble.com.
Cache.put latencies high
Hi, I have configured 2 ignite servers with a heap size of 8G each. Running with backups=1 and primary_sync mode. Ignite servers are being used as a write behind cache for my Aurora database. I am trying to run a load test with 3 clients talking to the remote cache in the 2 ignite servers. All the hosts are in the same availability zone. My clients do a simple put and I calculate time taken for put on the client side. long startTime = System.currentTimeMillis(); cache.put(k, c1); long elapsedTime = System.currentTimeMillis() - startTime; System.out.println("Total elapsed timein milliseconds: " + elapsedTime); Here is the latency from the last few requests: Total elapsed timein milliseconds: 31 Total elapsed timein milliseconds: 29 Total elapsed timein milliseconds: 26 Total elapsed timein milliseconds: 28 Total elapsed timein milliseconds: 28 Total elapsed timein milliseconds: 27 Total elapsed timein milliseconds: 27 Total elapsed timein milliseconds: 29 Total elapsed timein milliseconds: 26 Total elapsed timein milliseconds: 26 Total elapsed timein milliseconds: 27 Total elapsed timein milliseconds: 27 Total elapsed timein milliseconds: 28 Total elapsed timein milliseconds: 28 Total elapsed timein milliseconds: 27 Total elapsed timein milliseconds: 26 Total elapsed timein milliseconds: 27 Total elapsed timein milliseconds: 28 Total elapsed timein milliseconds: 27 Total elapsed timein milliseconds: 26 Total elapsed timein milliseconds: 26 Total elapsed timein milliseconds: 29 Total elapsed timein milliseconds: 26 Total elapsed timein milliseconds: 27 Total elapsed timein milliseconds: 26 Total elapsed timein milliseconds: 27 Total elapsed timein milliseconds: 29 Total elapsed timein milliseconds: 27 Total elapsed timein milliseconds: 28 Total elapsed timein milliseconds: 26 Total elapsed timein milliseconds: 27 Total elapsed timein milliseconds: 27 Total elapsed timein milliseconds: 27 Total elapsed timein milliseconds: 29 Not sure what is going on here. I am pretty sure I am doing something wrong here. Thanks!
Re: How is igfs:// URL resolved?
Denis Magda wrote > Hi Kobe, > > If I understood you task properly you want to map a path portion of an > HTTP request to an IGFS path. > > If this is your case then I would suggest doing the following: > - create a Web application that is running in a Tomcat container; > - the application will use Ignite in the embedded mode (start an Ignite > node from the app code); > - process incoming HTTP requests as usual extracting the path part from a > request; > - take instance of Igfs from previously started Ignite node and pass the > path and content to it. Refer to IgfsExample that is a part of Ignite > bundle/sources for more details. > > Regards, > Denis Denis, I went through the IGFS examples and did not quite see what I wanted. My problem is simply this: I am generating content into a file in IGFS that needs to be displayed on the browser. If the content (on the server) were in file webapps/myapp/content.gif I would return the URL http:///myapp/content.gif to the browser to the content may be rendered on the browser. Since my content is in IGFS, what would be the file:// URL of a file in (embedded) IGFS instance "myigfs" in path /data/content.gif as shown below? >IgniteFileSystem fs = ignite.fileSystem("myigfs"); > > // Image path. > IgfsPath workDir = new IgfsPath("/data/content.gif"); Thanx, kobe -- View this message in context: http://apache-ignite-users.70518.x6.nabble.com/How-is-igfs-URL-resolved-tp2142p2523.html Sent from the Apache Ignite Users mailing list archive at Nabble.com.
Re: Serialization issue with Ignite 1.5 built from master (357d791)
Alexey, Do you mean this? class MyWrapper { // Wrapped map. Map<> myMap = …; } D. On Tue, Jan 12, 2016 at 10:25 AM, Alexey Goncharuk < alexey.goncha...@gmail.com> wrote: > Paulo, > > Not sure if this will work for you as a workaround, but you can create a > thin wrapper around your map and put it into the cache instead. In this > case the map type is preserved. >
Cache read through with expiry policy
It seems that a read through does not set off an expiry policy. eg. cache = cache.withExpiryPolicy(new TouchedExpiryPolicy(new Duration(TimeUnit.SECONDS, 1))) cache.get(x) // triggers read through but this doesn't expire cache.get(x) // do "get" again. this will trigger expiry Can someone tell me if this is intended or not? Thanks -- View this message in context: http://apache-ignite-users.70518.x6.nabble.com/Cache-read-through-with-expiry-policy-tp2521.html Sent from the Apache Ignite Users mailing list archive at Nabble.com.
Re: Serialization issue with Ignite 1.5 built from master (357d791)
Paulo, Not sure if this will work for you as a workaround, but you can create a thin wrapper around your map and put it into the cache instead. In this case the map type is preserved.
Re: Node.js features
Hi, Node.js client will support the following features: - cache put/get operations; - queries (sql fields and scan); - task execution (run, affinityRun, mapReduce). I'm not sure about continuous queries in 1.6 release but I agree that it nice feature and probably it will be implemented later. You can track https://issues.apache.org/jira/browse/IGNITE-961 ticket in order to be aware of task progress. On Sat, Jan 9, 2016 at 7:08 AM, Dmitriy Setrakyan wrote: > Node.JS should be supported as part of 1.6. I am not sure about continuous > queries, but would be nice to have this supported. > > Andrey Gura, can you please comment here? > > D. > > On Thu, Jan 7, 2016 at 6:37 PM, ignite_mk wrote: > >> I see some activity on node.js integration as part of 1.6 release. >> Is there any plan to support continuous queries? >> Also, will there be support for messaging (queue) too? >> >> >> >> -- >> View this message in context: >> http://apache-ignite-users.70518.x6.nabble.com/Node-js-features-tp2441.html >> Sent from the Apache Ignite Users mailing list archive at Nabble.com. >> > > -- Andrey Gura GridGain Systems, Inc. www.gridgain.com
Re: Serialization issue with Ignite 1.5 built from master (357d791)
Dmitriy, For the lack of a workaround that will do it. But I can wait for a week or two before releasing my software. Do you think it will take more time than that? Thank you, Pires On Tue, Jan 12, 2016 at 6:09 PM, Dmitriy Setrakyan wrote: > Paulo, > > We probably won’t be able to release it quickly, but we can tag the fix > for you in GIT, and you can do your own build from it. Will that work? > > D. > > On Tue, Jan 12, 2016 at 9:10 AM, Paulo Pires wrote: > >> Hi Alexey, >> >> Is there any workaround I can use to circumvent this? >> >> If not, can you please release this as 1.5.1? Seems like a high/critical >> bug to me. >> >> Cheers, >> Pires >> >> On Tue, Jan 12, 2016 at 4:46 PM, Alexey Goncharuk < >> alexey.goncha...@gmail.com> wrote: >> >>> Thanks Paulo, >>> >>> I've reproduced the issue. Turned out the tests we have added were not >>> enough to cover all the cases. I fixed this behavior and now waiting for >>> the CI. I have reopened corresponding ticket: >>> https://issues.apache.org/jira/browse/IGNITE-2099 >>> >>> Let community decide when and how this fix should be released. >>> >> >> >> >> -- >> Cheers, >> Pires >> > > -- Cheers, Pires
Re: Serialization issue with Ignite 1.5 built from master (357d791)
Paulo, We probably won’t be able to release it quickly, but we can tag the fix for you in GIT, and you can do your own build from it. Will that work? D. On Tue, Jan 12, 2016 at 9:10 AM, Paulo Pires wrote: > Hi Alexey, > > Is there any workaround I can use to circumvent this? > > If not, can you please release this as 1.5.1? Seems like a high/critical > bug to me. > > Cheers, > Pires > > On Tue, Jan 12, 2016 at 4:46 PM, Alexey Goncharuk < > alexey.goncha...@gmail.com> wrote: > >> Thanks Paulo, >> >> I've reproduced the issue. Turned out the tests we have added were not >> enough to cover all the cases. I fixed this behavior and now waiting for >> the CI. I have reopened corresponding ticket: >> https://issues.apache.org/jira/browse/IGNITE-2099 >> >> Let community decide when and how this fix should be released. >> > > > > -- > Cheers, > Pires >
Re: Serialization issue with Ignite 1.5 built from master (357d791)
Hi Alexey, Is there any workaround I can use to circumvent this? If not, can you please release this as 1.5.1? Seems like a high/critical bug to me. Cheers, Pires On Tue, Jan 12, 2016 at 4:46 PM, Alexey Goncharuk < alexey.goncha...@gmail.com> wrote: > Thanks Paulo, > > I've reproduced the issue. Turned out the tests we have added were not > enough to cover all the cases. I fixed this behavior and now waiting for > the CI. I have reopened corresponding ticket: > https://issues.apache.org/jira/browse/IGNITE-2099 > > Let community decide when and how this fix should be released. > -- Cheers, Pires
Re: Serialization issue with Ignite 1.5 built from master (357d791)
Thanks Paulo, I've reproduced the issue. Turned out the tests we have added were not enough to cover all the cases. I fixed this behavior and now waiting for the CI. I have reopened corresponding ticket: https://issues.apache.org/jira/browse/IGNITE-2099 Let community decide when and how this fix should be released.
Re: SQL Query returns no value if different values put twice with same key
Hi Val, You are right. I totally missed the point that i changed id of 2nd object. There was no system in my network and visor also showed nothing. I will restart system and try again. How about class not found exception. Am i doing something wrong in this code which should cause this exception on server node? Regards, Vinay
Re: Ignite Events overhead
Hi, Please see my answers inline On 1/12/2016 10:23 AM, pavlinсм wrote: Hello, I have two questions about events : 1) In the documentation one can read "/Since thousands of events per second are generated, it creates an additional load on the system./" What kind of load to expect with local events - network, CPU, memory, overall ? In case of local events both CPU & memory are consumed. CPU is used because an event object has to be constructed and processed by a local listener and stored in EventStorageSpi [1]. Additional memory is occupied by created event objects and EventStorageSpi implementation. 2) If remote events are enabled will there be additional network traffic between nodes ? Yes, if a remote filter succeeds then an event is delivered to a local node's local listener that started the remote filter. [1] https://ignite.apache.org/releases/1.5.0.final/javadoc/org/apache/ignite/spi/eventstorage/EventStorageSpi.html -- Denis -- View this message in context: http://apache-ignite-users.70518.x6.nabble.com/Ignite-Events-overhead-tp2511.html Sent from the Apache Ignite Users mailing list archive at Nabble.com.