@Audrey, can you provide me a snippet how can I override ignite api's testing
implementation?
--
View this message in context:
http://apache-ignite-users.70518.x6.nabble.com/Class-objects-are-fetched-as-string-when-JDBC-api-is-used-tp8720p8772.html
Sent from the Apache Ignite Users mailing list
I went to local gradle repo and manually deleted other h2 versions. It’s
working now.
I have one more query – is there a way I can make ‘ignite-http-rest’ work
with spring-boot? Last time when I checked with one of Ignite dev they
mentioned that I can’t use it with boot. Without rest-api, I need t
Hi,
I am looking at a feasibility of using Ignite as a persistence layer
instead of a mySql/Postgres db where we do lot of processing before sending
data to our rest-api.
1. Is it good to use ignite as a storage?
2. Is it efficient to do so much processing of data in ignite?
3. What is the avail
Hi Roman,
Thank you for the brief explanation, I will try to modify the source code of
FileStreamSourceTask.java
Thanks & Regards,
Austin
--
View this message in context:
http://apache-ignite-users.70518.x6.nabble.com/Kafka-Failed-to-stream-a-record-with-null-key-tp8731p8776.html
Sent from th
Hi,
There are no such API in C++ API yet. We are working on
extending C++ API though so it may appear in future.
Best Regards,
Igor
On Tue, Nov 8, 2016 at 10:41 AM, smile wrote:
> Hi, all
>I used ignite C++ API, and I find that it dose not support pub/sub ,
> also not support event, so is
Hi,
Yes, you are right, if you decrease the property (sqlOnheapRowCacheSize),
you get sacrificing performance.
But default value is really great:
public static final int DFLT_SQL_ONHEAP_ROW_CACHE_SIZE = 10 * 1024;
You can try to analyse heap dump, in order to make sure in the reason of
the iss
Hi,
I've checked - by some reason this file really is missing from the release
package.
It is present in the 1.7.0 tag in git repo though. You can get it here -
[1]
[1] - https://github.com/apache/ignite/releases/tag/1.7.0
Best Regards,
Igor
On Tue, Nov 8, 2016 at 6:45 AM, smile wrote:
> Hi,
Hi Shawn.
It looks strange that each entry costs 10k in average, but you expect 50
bytes at most.
You wrote that there are 300k entries, but I see 5000k keys in Heap.
Grid configuration and Heapdump would be helpful to understand who holds
all that data. Or It would be great to have a reproducer
Hi Tracyl,
Can you describe in greater detail what you are trying to achieve? To my
knowledge, predicate pushdown is a term usually used for map-reduce jobs.
The concept of Ignite's jobs and tasks is more similar to fork-join rather
than map-reduce semantics, so we could better help you if you des
Hi Andry,
It is look like a mistake:
cache.put(1, new TestVal("old"));
TestVal oldVal = cache.get(1);
oldVal.val = "new";
cache.put(1, oldVal);
You need create new object always.
Try to do like this:
cache.put(1, new TestVal("old"));
cache.put(1, new TestVal("new"));
Otherwise you can to get s
Hi,
can you help me how to perform this query with Ignite SQL.
I have this class
class ClientUpdate
{
private final String registrationId;
private final LinkObject[] objectLinks;
}
class LinkObject implements Serializable{
private final Integer objectId;
private
Hi Vladislav,
Thanks for the reply.
Example I have provided just to explain our case. In our system we wanted to
update several properties on existing objet in the cache but don't make full
copy of object each time we extracting object from cache.
We are using default configuration for copyOnRea
Hi Val
Thank you for responding to my post. I agree with the idea of using
Ignite for both read/write through cache.
However in my application, writes to persistent store happens through
a different application.
My application is a reporting application, where I need to stream the
data, apply com
Srini,
In this case you will have to implement a mechanism on the DB side that will
use Ignite API to update the cache. Ignite doean't provide anything out of
the box for this.
Another way to handle this is to configure expirations for cache entries
[1]. This way any entry will be eventually evic
Thanks Alexey.
By predicate/projection pushdown, I mean: currently I am storing a native
Spark Row object as value format of IgniteCache. If I retrieve it as an
IgniteRDD, I only want certain column of that Row object rather than
returning entire Row and do filter/projection at Spark level. Do you
Devis,
Having a link from one object to another is the only way to join them. You
need to add a 'clientUpdateId' field to LinkObject, store each LinkObject as
a separate entry and then join. Another way is to denormalize and add
'registrationId' to LinkObject and remove ClientUpdate type. This way
Hi,
Please properly subscribe to the mailing list so that the community can
receive email notifications for your messages. To subscribe, send empty
email to user-subscr...@ignite.apache.org and follow simple instructions in
the reply.
alex wrote
> When I run the code below, I got "Failed to peer
Hi,
Please properly subscribe to the mailing list so that the community can
receive email notifications for your messages. To subscribe, send empty
email to user-subscr...@ignite.apache.org and follow simple instructions in
the reply.
Navneet Kumar wrote
> I am connected with the remote ignite i
regarding the - configure expirations for cache entries.
what happens when the cache is expired ? underlying query will be executed
again by the data streamer and data will be loaded into cache again ? please
clarify.
--
View this message in context:
http://apache-ignite-users.70518.x6.nabble.
Hi,
It's fine to use Ignite as the main and only data storage for your
application, but Ignite is not a persistence storage. Data is in memory, so
there is always a chance for data loss. If this is something that you can't
live with, then do not rip and replace, but use Ignite with a persistence
s
Hi,
If there is a high load, I think there is a big chance of losing something
in 10 seconds. However, you can try to increase flushSize property which
controls the maximum amount of entries saved in queue. Note that in case you
can't tolerate any data losses, you should use write-through instead
Hi,
You get both old and new value in the event. So you can compare this field
values within your listener and act accordingly.
-Val
--
View this message in context:
http://apache-ignite-users.70518.x6.nabble.com/Fwd-How-to-trigger-out-event-when-a-specific-field-got-updated-tp8767p8797.html
Hi,
I tried to run your code and it works properly for me. Are you sure you're
running the exact same version?
Also note that you must do a full copy of the object before putting it back
to cache. If you try to avoid this, you can get unexpected behavior because
you will mutate the actual instanc
To my knowledge, both JDBC driver and especially SqlFieldsQuery should return
objects, not strings. However, if you print out these object and you didn't
override toString, then you will get exactly what you shown in the output.
Can this be the case?
-Val
--
View this message in context:
http:
Hi,
It looks like cache object is unexpectedly serialized. Is it referenced by
loadCache predicate and/or arguments?
-Val
--
View this message in context:
http://apache-ignite-users.70518.x6.nabble.com/getting-Failed-to-find-cache-even-though-cache-is-there-tp8787p8800.html
Sent from the Apa
If these values are read-only in this transaction, why do you need to lock
them in the first place? Shy not to read out of transaction, and then use
values inside the transaction?
-Val
--
View this message in context:
http://apache-ignite-users.70518.x6.nabble.com/Explicit-lock-whithin-a-trans
When expiration happens, the entry is removed from cache. Once you try to
read it again, it will be reloaded.
I'm not sure how you use data streamer though. It actually updates the
cache, but you just told that cache is read-only and updates are going
directly to DB. Can you clarify?
-Val
--
V
Hi,
This happens because you execute query on the client. When you're querying a
local cache it never goes to a remote node, but client doesn't have any data
in your case. I think there should be a proper exception in this case
though.
-Val
--
View this message in context:
http://apache-ignit
thanks Val for quick response.
1. Regarding 'read-only cache', below is code snippet from my cache config.
As my application is a reporting application, the writes to database will
not happen via my cache.
hence i set 'setWriteThrough' to false and set 'setReadThrough' to true.
Srini,
Periodical reload is also an option. But you will need to be able to know if
the row in the database was updated since last reload. E.g. an 'updated'
column in the table or something like that. Otherwise you will have to
reload everything, because out of the box Ignite is not aware of any u
Thanks very much for replying, @vkulichenko
I run this code with a junit test case in IDE.
server side is alive, test case setClientMode(true).
And I had tried to package the project and run it in terminal, I got the
same Exception. After Exception was throw out, Client and Server not exit in
t
When I run the code below, I got "Failed to peer load class".
String objectIds =
"1vxzn3ifggm4o,1a47fmqipb1u3,z56f5kkwlfk3,tths3z5k5l38,79lzqlrd4cg6";
for (String did : objectIds.split(",")) {
calls.add(() -> {
TLabObject object = RepositoryDao.getObject(did);
System.ou
One more query Val, Is Node.js support for Ignite (client and server) available?
--
Regards,
Chetan.
From: "vkulichenko [via Apache Ignite Users]"
Date: Wednesday, November 9, 2016 at 3:31 AM
To: "Chetan.V.Yadav"
Subject: [EXTERNAL] Re: Query on using Ignite as persistence data and
processing
I am not printing it out directly but converting resultset to list and passing
to my rest-api. Overriding toString() provides me data on first level variables
say class A but I have other class objects (say B) list inside A which is
returning empty list even if I override toString() inside class
Duplicated my answer:
If you have 2 backups (in case you have enough nodes) this mean you can
loose 2 node, without afraid to loose data. In case you have a fail more
than the 2 nodes, it is mean you have incomplete(only part of data has
available) data.
In general, If you are loosing more nodes
35 matches
Mail list logo