Hi, thanks for clarification.
My use case it that I have some logic that is hidden from client but client
can execute this logic by calling Ignite Service.
Some of those operation may take a while, so I thought that it would be nice
to run it asynchronously and return Future object so the client
It is not possible to start the communication/discovery service on public ip. I
get a BindException: Cannot assign requested address whenever I do that.
Therefore, I was starting the services on Private IPs.
From: Denis Mekhanikov [mailto:dmekhani...@gmail.com]
Sent: Monday, November 27, 2017 10
In order to use all the resources in a single machine, I want to deploy
multiple service instances in a single machine, and incoming service
requests to be spread over these instances.
However, according to the documentation, it seems the load balancing only
works in a cluster node level, not a se
Client node driver mode could bring a huge memory usage on your application (
local cache)
thus I believe most choose JDBC thin driver mode.
BTW our data source library called Druid doesn't support common JDBC Driver,
which reported transaction issues.
So, this feature seems very important for
That's exactly what I expected
--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/
You can use client node based driver [1] to get full failover support. Thin
client currently indeed goes though a single node and I believe it makes
sense to at least provide an ability to specify multiple addresses for the
connection. I created a ticket for this:
https://issues.apache.org/jira/bro
Denis,
For this cross region case, I've some thought to share and please help
verify whether they are already here.
1. Two DC, A ---corss region--- B
2. A have several ignite node, A1,A2,A3 and B have several ignite
node, B1,B2,B3 ..., there are in a same cluster
3. In ingite, A1,A2, A3
This improvement makes sense, but I doubt there are any plans to implement it
at the moment. Feel free to create a ticket in Jira.
-Val
--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/
Alexey,
Something is wrong, but I don't see any obvious mistakes in your code. Is it
possible to provide a test as a standalone GitHub project so that I can run
it and reproduce the problem?
Is it reproduced on smaller data sets? Or if load not through Spark, but
just doing regular put/putAll ope
Dennis,
Sorry for my misleading. The slow thing is register CQ but not connect to
cluster (this might take round 1 min).
Thanks for your remind and after removed intial query, the whole process
become very quick and there is no more "binary metadata update messages".
So if there do have initia
Krzysztof,
This will no work because service is invoked in server side and the returned
future gets serialized and sent to client. The deserialized instance is
obviously never completed. What are you trying to achieve?
-Val
--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/
Use case is discussed at length:
http://apache-ignite-users.70518.x6.nabble.com/Continuous-update-Data-Grid-Cache-td2075.html#a17641
Currently, we have to load data to second table, the client then has to
change their query to query this
new table. Drop the old table.
The latest suggestion in t
Hi Wolfram,
Looks like it's broken, we parse those type bits in a request, but don't set
in for a response.
I created a ticket for this problem:
https://issues.apache.org/jira/browse/IGNITE-7028
Thanks,
Mike.
--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/
It appears Ignite has load balancing support for its compute grid but not for
data grid.
As in, the JDBC thin client does not have load balancing and failover
support.
If JIRA for it does not exist yet, can we create a JIRA for Ignite JDBC thin
client to provide load balancing and failover suppor
Hi Team,
I have a question regarding Ignite Services and Compute async.
If we have a deployed IgniteService like this
public class MySercviceImpl implements Service, MyService {
@IgniteInstanceResource
private Ignite ignite;
@Override
public IgniteFuture doStuff() {
IgniteFutu
Hi,
I've loaded 50 million BinaryObjects into TEST cache using Apache Spark
They look like this:
o.a.i.i.binary.BinaryObjectImpl | DATASET1 [hash=86282065, F01=-206809353,
F00=A1782096681-B2022047863-C554782990, F03=Must be timestamp,
F02=2.6983596317719E8, F05=182918247,
F04=A1997114384-B293944
Hi Denis,
Thank you for confirming the issue. Is there any other way to calculate cache
size? Also, how accurate are the other cache metrics like
getAverageGetTime/getAveragePutTime? Do they provide average get time for
entire cluster or just local node?
-Biren
From: Denis Mekhanikov
Reply-T
Wolfram,
The buffer size is hardcoded now, but it could be made configurable if it
the real issue.
Can you share a simple PHP sample?
I think Java Unit tests may miss some important details you have with PHP. I
wonder if I can test & find the missing part of the puzzle ).
Thanks,
Alexey
--
Se
Hello!
The recommendation here is running two nodes from two different
directories: such as apache-ignite-fabric-2.3.0-node1 and apache-
ignite-fabric-2.3.0-node2.
Otherwise such collisions may occur. Logs for two instances might also get
clobbered when run from the same directory.
Regards,
--
Hi Rajarshi,
I see that you use KafkaProducer in you compute job.
Are you sure that KafkaProducer can be safely serialized on one host and
deserialized on a remote host and deserialization version can be safely on
the remote host?
I think you need properly create/obtain KafkaProducer on a remote
Hi,
I have found I regularly get exceptions such as the one below if I have 2
nodes on the same server trying to run IgniteCallable jobs at the same time.
Can anyone give me a hint on how to prevent this - it can prevent a job from
completing successfully.
Thanks
[16:45:27,217][SEVERE][pub-#68%n
Hi,
It seems misconfiguration. Could you share your cache configuration and
double check that you set CacheConfiguration#setAtomicityMode to
TRANSACTIONAL instead of ATOMIC which used by default?
On Sat, Nov 25, 2017 at 9:19 AM, Sumanta Ghosh
wrote:
> Hi,
> I am using Ignite's SpringTransactio
Hi,
Because our company's network do not allow send file to the internet,so I
try to decribe the steps:
1.I got all records from two tables, it's very simple,just like this:
select a.id,a.name,b.depatment from tablea a inner join tableb b on a.id=b.did
where a.id in('a','b') ; this step
Sorry, I still don't understand, why you need an AddressResolver.
Can't you specify all external IPs in configuration and don't use internal
ones?
Denis
пн, 27 нояб. 2017 г. в 15:55, Josephine Barboza :
> Hi Denis,
>
>
>
> I was able to establish connection between the two nodes. The problem was
And are you sure, that it's not initial query causing continuous query to
register slowly?
Maybe class loading has nothing to do with it?
Denis
пн, 27 нояб. 2017 г. в 19:26, Denis Mekhanikov :
> You wrote in your previous letters, that startup takes much time.
> Could you clarify, whether node c
You wrote in your previous letters, that startup takes much time.
Could you clarify, whether node connection to a cluster takes much time, or
query execution? Because I'm a bit confused.
The messages, that you see is not a class deserialization, it's binary
metadata update messages. They appear wh
I don't see much sense in using both HDFS and Ignite native persistence,
since HDFS stores data on disk on its own.
Could you describe your use-case?
Maybe, there is a point to allow it, if there is a valid usage scenario.
Denis
сб, 25 нояб. 2017 г. в 6:31, Amol :
> Thanks for quick reply. Any
Hi,
Another simple check:
How many thread do you use for update & select? In case your application
is multi-threaded be sure that separate JDBC Connection is used for each
thread. Ignite JDBC API is not thread safe.
Otherwise, please share the smallest reproducer / steps to reproduce.
On 27
Could you attach queries, that you are trying to execute and full error
message?
Also make sure, that JDBC driver version is also 2.3
Denis
сб, 25 нояб. 2017 г. в 12:39, Lucky :
> Hi
> The ignite version is 2.3
> When I update something by cache.query(new SQLFieldsQuery(sql)), and
> the
Hi,
Apparently the very same bug also affects the standalone kafka connect:
When I add multiple connectors it will start the last one given, shut that
one down (resulting in another 'Data streamer has been closed' exception)
and then start the next one - So there's always only a single one running
Hi Hyma!
Looks like you encountered a classic deadlock. It happens because you put
values into cache in arbitrary order.
This line causes this problem:
*companyDao.nameCache.putAll(kvs)*
So, when multiple threads try to acquire the same locks in different order,
then these operations will be wait
Hi!
Thanks for your answer, great insights.
It would be great if you could open a Bug for that. I'll have to go with
the standalone version for now either way.
I will try to implement a Kafka converter for that as you suggested.
Let's see if I can get it running!
Best Regards
Svonn
Am 21.11
Hello Alexey,
I'm currently running OpenJDK 1.8 on Linux (4.9.65-1-MANJARO):
openjdk version "1.8.0_144"
OpenJDK Runtime Environment (build 1.8.0_144-b01)
OpenJDK 64-Bit Server VM (build 25.144-b01, mixed mode)
I tested it with a PHP application, because I wanted to connect a legacy
service whi
Has anyone use Ignite Native Persistence with Docker?
Is there a solution on how to map the Volume dynamically? And how about when
you restart the whole cluster, how does it maps all volumes?
--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/
Hi Denis,
I was able to establish connection between the two nodes. The problem was that
in AddressReslover the internal-external address mapping was incorrect. I had
to specify the socketAddress instead of the address as the internal address for
it to work.
addrs=[10.131.12.16], sockAddrs=[/10
Hi Wolfram,
I just run unit tests for Redis with 10k string. They passed without errors.
Can you share a reproducible example?
Actually, the issue happens inside java.nio.HeapByteBuffer. What jdk/jre
version do you use?
BTW, this buffer size is set to 8k and it should be re-used in your case.
T
Hi Mikael!
You don't need to check whether a service is already deployed.
Ignite will handle such collisions, but only if you don't try to deploy the
same service multiple times with different configurations. Otherwise you
will get an exception.
If configurations match, then consecutive deployment
Hi!
I have a number of services running (cluster singeltons) and each node
has java code to start them up (they are configured from information in
a database), so each node tries to start the services when they are
started, should I check if the service is already running or something
or will
Hi!
You shouldn't set thread pool size to such high values.
It's recommended to leave it default (number of available cores), or 2 x
number of cores.
> I found Public thread pool idle numbers on every node is always 0
What do you mean by that? How did you check it?
CPU load may be low, if tasks
Thanks for a fast response :)
All is clear now.
Regards,
Krzysztof
--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/
Hi,
You do not need to always create a service proxy. You can get a proxy once
and it will always find a service even if the current node where the
service is deployed failed: the service would be re-deployed and the
service proxy would find it on the new node. That is true even if you
created a "
Hi Team,
I would like to ask about a preferable approach regarding the Service
acquiring via Ignite's Service Proxy and a node failure.
Lets assume we have a service deployed as a Cluster Singleton. Grid client
acquires the service instance via ignite.services().serviceProxy(...)
The questions a
Normally when storing data in Ignite using the default
RendezVousAffinityFunction, data is distributed reasonably evenly over the
available nodes. When increasing the cluster size (up to 24 nodes in my
case), I'm finding that the data is not so well distributed. Some nodes have
more than twice as m
Mikael,
If you choose to use *AffinityKey* to wrap your keys, then you should use
it everywhere. Because basically it's just a tuple of a key and affinity
key, so this pair is used as a composite key.
You can also annotate some field of you key class with *@AffinityKeyMapped*,
instead of using *A
Hi,Denis:
2.3 will be better than 2.2.
But the performance improvement is not obvious.
It took 2.5 seconds in single ignite node ,and took 5 seconds in cluster(3
ignite nodes).
JDBC Client driver is worse than JDBC thin driver. it took about more 1
seconds.
Did there a
Thank you dear ALexey,it wors for me. )
--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/
Hi Daniels,
Actually, you can have & use indexes for single (1 to 1 relation) nested
objects. You can't have indexes for nested collections.
"Indexes for nested *collections* and in particular for *maps* are not
supported.".
I wonder why Examples does not have that. Please see how it could be
47 matches
Mail list logo
