Re: Eviction policy enablement leads to ignite cluster blocking, it does not use pages from freelist

[ezhuravlev]

Hi,

>From logs, I see that all server nodes are pinging node on host 10.180.48.14
all the time. It can be caused by the lack of the connection between part of
the nodes(for example, if node 10.180.48.14 can't directly communicate with
client nodes). I think that client nodes might ask server nodes to check if
node on 10.180.48.14 is still alive. I would recommend checking connectivity
between all the machines in the cluster, including client machines.

Best Regards,
Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Cache Expiry policy not working..

[ezhuravlev]

Can you please share the entire source code? We also need to see how you're
checking that data is still there.

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Persistent dataregion config

[ezhuravlev]

Hi Andrey,

Sorry, Looks like mixed up eviction policy for data region and Expiry
Policy. Eviction Policy shouldn't work with persistence, it works only for
in memory data regions.

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Offheap memory consumption + enabled persistence

[ezhuravlev]

>But I am still confused. Did I understood properly that each node should 
have enough memory to store full data set? 
No, it's not the case. It's just preallocating this memory with internal
structures and shows it in the metrics.



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Offheap memory consumption + enabled persistence

[ezhuravlev]

Hi,

Growth of this metric will be stopped after all partitions(1024 by default)
allocate memory for maximum count of entries that can be stored in it
simultaneously. It will take some time for 1024 partitions(not 100
iterations). You can check it by setting partitions count to smaller size:
cacheConfig.setAffinity(new RendezvousAffinityFunction(false, 8));

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Serialisation issue.

[ezhuravlev]

Hi,

I'm able to start the cache with this configuration. Which version do you
use? Can you share small project with a reproducer, as this code snippet
works without a problem?

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Improving Get operation performance

[ezhuravlev]

As you have 4 nodes on the same machine now, you have a lot of context
switching, probably all the nodes just competing for CPU resources with each
other.

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Server Nodes Stopped Unexpectedly

[ezhuravlev]

Hi,

Answers to questions from the previous message will be based on the provided
logs, since it's not clear what happened there yet.

IgniteConfiguration.setNetworkTimeout: 
It is a global timeout for high-level operations where a network is
involved.

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

RE: Ignite bulk data load issue

[ezhuravlev]

This message:
 Blocked system-critical thread has been detected. This can lead to
cluster-wide undefined behaviour [threadName=data-streamer-stripe-0,
blockedFor=34s]
is being printed in case if you're loading data to Ignite much faster than
it can write to the disk. The disk is a bottleneck here, as you mentioned
before, you chose a disk with a really small IOPS. Just give a better disk
to the system and you will see a better performance.

By the way, how much heap do you have for a JVM?

Evgenii 



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Running 2.x Ignite in embedded mode with persistence enabled

[ezhuravlev]

Hi,

> already running on high heap space (up to 10GB)
I wouldn't say that 10GB is a high heap usage, so, it shouldn't be a
problem.

> It’s on windows VMs, with HDD disks, and by-chance some windows VMs  could
> run multiple instances of this application. 
More likely running multiple Ignite nodes on one machine with HDD will be a
bottleneck for application - these nodes will compete for the I/O resources.
Why did you decide to choose this architecture? Did you consider running an
Ignite cluster separately from application and just having client nodes
embedded in the application? What do you use there at the moment?

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: critical thread has been detected

[ezhuravlev]

Hi,

Unfortunately, I can't see any exception message, could you please try to
attach the log file?

Thanks,
Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Query execution too long even after providing index

[ezhuravlev]

How much data do you have?  What is the amount of heap and offheap memory?
Can you share the reproducer with the community?

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: ./control.sh --baseline doesn't list nodes on a linux machines

[ezhuravlev]

>By the way - I wonder what is the right way to reset an Ignite instance to 
the initial state (the one I have when Ignite is just installed)? 
You should delete work and db directories.

>I could not activate the new cluster. When control.sh 
--baseline tried to connect to the cluster it hung 

Can you share the logs from the nodes?

I'd recommend to check all processes on the machine - topology version
you've mentioned could be caused by the running visor somewhere.

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Service Grid classpath changes

[ezhuravlev]

When you deploy the service, it serializes the object of this service and
puts it to the internal cache. At the moment when it should be redeployed,
it will take this object from the cache and deserialize it.

So, if after change your object still can be deserialized using the new
version of the class, you can change it. But in this case, it's possible
that you will have the different versions of the class on the different
nodes.

If after the change Service object can't be deserialized, you can just
deploy it with the new name, use it for a short time, undeploy the old one
and deploy it with the new version of the class again.

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Unable to get the ignite cache metrics

[ezhuravlev]

Hi,

Looks like you try to access local cache metrics on client node, which will
be empty, because clients doesn't store cache. Try to use
CacheClusterMetricsMXBeanImpl instead.

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Using compute().call from a Service has P2P class loading feature?

[ezhuravlev]

Hi,

Are you sure that you use proper service? I mean from the client node. I've
just checked this case and everything works fine.

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: WARN TcpDiscoverySpi:135 - Timed out waiting for message to be read(most probably....)

[ezhuravlev]

Hi,

Do you have connectivity between nodes? Have you checked that all nodes are
reachable and there are no long GC pauses?

Can you share the full logs from all nodes?

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Ignite Visor cache not active

[ezhuravlev]

Hi,

Can you show the topology which Visor prints?

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: ./control.sh --baseline doesn't list nodes on a linux machines

[ezhuravlev]

If you just started 3 nodes and you have topology version 20, then, probably,
you have started visor started somewhere, because even daemon nodes can
store topology version.

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Does ignite supported calculated columns ?

[ezhuravlev]

Hi,

How often do you access it? Maybe it would be better to calculate it each
time you access it in getter method?

Alternatively, you can implement Binarylizable interface and calculate this
column there, before serializing it.

Evgenii 



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Query execution too long even after providing index

[ezhuravlev]

Hi,

What is the execution time of just a single query without UNION?

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: 答复: why my ignite cluster went to compatibility mode

[ezhuravlev]

Hi,
>but why one of my ignite nodes went down abnormally?
Can you share logs with the community so somebody could take a look into it?

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Proper config for IGFS eviction

[ezhuravlev]

Hi,

There are metaCacheCfg and dataCacheCfg in
IgniteConfiguration.FileSystemConfiguration, which allows to configure
caches for IGFS. Inside this configuration you can set DataRegion with
persistence.

Have you tried this?

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: ignite multi thread stop bug

[ezhuravlev]

Hi, 

it's a really bad idea to do cache operations in Discovery thread. If you
want to add something to the cache, you can start a new thread and do this
operation inside the new thread.

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Data migration using persistent Queues

[ezhuravlev]

Sorry, here is the link for the BinaryObject doc:
https://apacheignite.readme.io/docs/binary-marshaller



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Data migration using persistent Queues

[ezhuravlev]

Hi,

Do you mean Ignite Queue? Or just your data structure that stored in simple
Ignite cache?

> Store data in another format let's say JSON and provide backwards
> compatibility at code level.
There is no need for this, Ignite internally stores everything as
BinaryObject and this format supports schema changes: Store data in another
format let's say JSON and provide backwards 
compatibility at code level.



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Load balancing ignite get requests

[ezhuravlev]

Hi,

Well, it depends on a lot of things - if you have small amount of the data,
which can easily fit in memory on each node, then, you can use Replicated
cache.

On the other hand, if you have quite big dataset, you may consider using
Partitioned cache and executing affinity runs.

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: The system cache size was slowly increased

[ezhuravlev]

Hi,

What do you mean by " system memory cache also grows "? How do you see this?

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Can't write to Ignite cluster when one node in baseline is down

[ezhuravlev]

Hi,

It's ok for the case when you run without backups, because you lose
partitions here.

Do you start a new cache each time or reuse the cache from the different
run? When you use persistence, Ignite persists cache's metadata on disk and
if you change cache configuration after that, these changes won't be
applied. I mean, it's possible that when you set backups to 1, after that
you still run cache with old metadata, without backups.

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Hibernate L2 cache with Ignite. How?

[ezhuravlev]

Hi,

Are you sure that you access the same cache that is used for l2? Try to
invoke Ignite.cacheNames() - it will give you a list of caches in the
cluster, using that, you can check that use right cache.

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: SqlQueryFileds on List of Binary Object

[ezhuravlev]

Hi,

As for now, it's not possible to query objects if they're stored as
collection like IgniteCache
>, so, I'd recommend changing the way how you
store objects in cache - it will be pretty the same if you will store it
like IgniteCache

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: service deploy cluster singleton lock application

[ezhuravlev]

Hi,

As per java doc of DataRegionConfiguration.setMaxSize - maxSize Maximum data
region size in bytes. For sure, it won't work properly with such a small
memory size(128 and 1024 bytes). Ignite even have validation for this
parameter and it should be not less than 10mb at least.

Have you checked your logs? Do you have any exceptions? Have you checked
thread dumps? Where was it hanged?

Regards,
Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: If the authenticationEnabled is true Ignite web control conneced failed

[ezhuravlev]

As for now, web agent can't connect to the secure Ignite cluster, this
functionality should be available with version 2.7

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Call to deployClusterSingleton blocked.

[ezhuravlev]

Hi,

Is this issue reproducible? Can you share logs from all nodes and code of
the service?

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Hive (IGFS + IgniteMR) vs Hive (Tez)

[ezhuravlev]

Hi,

1. Have you tried to run it without -Dignite.job.shared.classloader=false ?
It definitely has a performance impact.

2. Are Ignite nodes placed on the same machines as Hadoop? If not, it will
add a huge network interaction.

3. What is the amount of the data that you have in hdfs? If it's not fit in
IGFS(as I see, you have something like 90gb in IGFS), you will have a lot of
data moving between memory and hdfs - it will remove old data on reading a
new one every time, when there is not enough memory.

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Failsafe re-entrant locks still held after a node exits.

[ezhuravlev]

Hi,

I've checked code and it looks like the current implementation of
ReentrantLock doesn't have proper processing of failoverSafe flag. I've
created ticket for it's implementation:
https://issues.apache.org/jira/browse/IGNITE-8810.

As for workaround, you can use reentrantLock without failoverSafe flag and
isBroken() before locking and recreate lock if needed. Alternatively, you
can add listener for Node_LEFT and NODE_FAILED events and check locks on it.

Additionally, you can use CacheLock instead of ReentrantLock - it works even
faster: https://apacheignite.readme.io/docs/distributed-locks

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Ignite latest release using zookeeper 3.4.6

[ezhuravlev]

Hi,

I've created a ticket or zookeper version upgrade:
https://issues.apache.org/jira/browse/IGNITE-8433. 

Thanks,
Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Intermittent Spikes in Response Time

[ezhuravlev]

Hi Chris,

How do you map compute tasks to nodes?

Well, it's possible that 2 nodes in your cluster always stores more data
than others, that's why you face these spikes, it could happen in case of by
your affinity key, too much data could be collocated. You can check it by
using IgniteCache.localSize on each node and comparing results.

Also, I would recommend checking Ignite.affinity.mapPartitionsToNodes - just
to make sure that all nodes store pretty the same amount of partitions.

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Ignite cache get operation not working, client application stucks at the initializaiton

[ezhuravlev]

You need to find field "ht" in your objects.

It's needed to have the same files for both client and server nodes.

I've asked in the previous message to provide logs from client nodes, could
you please share it?

Thanks,
Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Did update SQL support joiner?

[ezhuravlev]

Hi,

First of all, you need to check that all data is collocated since by default
join will be performed on each node on the collocated data.

Also, I don't see FROM in UPDATE grammar in H2:
http://www.h2database.com/html/grammar.html#update
So, I would recommend to rewrite it using subquery.

Regards,
Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: How to achieve writethrough with ignite data streamer

[ezhuravlev]

if skipStore flag is enabled, then data won't be propagated to the store


are you sure that you have writeThrough flag enabled?

Additionally, you need to check your CacheStore implementation, if you don't
have any ideas, you can post it here, the community will check it.

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Understanding data store and partitioning

[ezhuravlev]

Hi,

1) You can somehow signal it from your application where you invoke
loadCache method or from CacheStore implementation 

2) loadCache is sync method and when it will be finished, all preloading
will be finished on all nodes

3) You can listen to the topology changes and check count of the nodes in
topology: https://apacheignite.readme.io/docs/events

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: The wrong sockAddrs are registered, and the cluster is broken when it tries to connect it occasionally.

[ezhuravlev]

If it really starts to use addresses other that pre-configured, you should
share your configurations and full logs from all nodes, the community will
check it.

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

RE: CacheStoreAdapter not getting called on cache operation.

[ezhuravlev]

Hey Pim,

Looks like your question was already answered here:
http://apache-ignite-users.70518.x6.nabble.com/CacheStoreAdapter-write-and-delete-are-not-being-called-by-Ignite-s-GridCacheStoreManager-td19624.html

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Re: Cannot connect the ignite server after running one or two days

[ezhuravlev]

 >There are several clients connected by VPN,  is it possible to the client's
restart regularly causing ignite socket communication to a certain degree of
obstruction and becoming more and more serious as time goes by? 

Does this mean that it's impossible to connect from server to client nodes
directly? If so, it can definitely be the cause of this behavior.



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Re: Cannot connect the ignite server after running one or two days

[ezhuravlev]

Hi,

Looks like logs from the server is still not full. If you've checked them
and you sure that you don't have any exceptions in it before witnessing this
problem, then, I think that you could have some connection problems or a
long GC pause. Do you have any network monitoring? Also, I would recommend
checking GC logs at this moment(or share it with community) on all nodes.

Also, Why did you set "queryThreadPoolSize" value="32" while you have only
24 CPUs for all 5 hosts? It's definitely will reduce performance due to a
lot of context-switching.

If everything okay with GC logs, it's possible to check TCP dumps, just to
understand where the connection breaks off.

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: No transaction is currently active || Not allowed to create transaction on shared EntityManager - use Spring transactions or EJB CMT instead

[ezhuravlev]

Hi,

I think now you need to start and commit transaction manually in your code:
em.getTransaction().begin();
em.getTransaction().commit();

At least I've tried it some time ago and it worked. Please check this
example on github:
https://github.com/ezhuravl/ignite-eclipselink-example

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: horizontal scaling

[ezhuravlev]

Hi,

> Right now we are running 4 nodes (same physical system)

Do you mean that you run 4 nodes on the same physical machine and after
this, you add 2 additional nodes on the same physical machine?

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: ScanQuery with predicate always returns empty result.

[ezhuravlev]

Hi, in case when you use:
ScanQuery scan = new ScanQuery<>((key, value) -> {

System.out.println(key + “ = “ + value);

return true;

})
I think the problem is that you created ScanQuery while
your cache is IgniteCache. Scan query should be  too.

In case of Inner class, it's trying to serialize your outer class, but it
faces some problems in this process. Here you can create separate class
instead of inner, it will help

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Native persistence

[ezhuravlev]

Hi Humphrey,

>What will happen if at a later point I want to scale back up to 4
replica's? 
>- So what will happen with the data it finds in the existing directory (and 
probably are old), how does it handle this? 

It depends on the time when the node was down. If it was a short period of
time and it's possible to apply delta rebalance - the current partitions
will be updated to the last version. Otherwise, old partitions will be
dropped and full rebalance will happen, like for new nodes. The possibility
of applying delta rebalance depends on configured walHistorySize.

>- What happens in the situation that I shutdown my cluster and restart it 
with 2 replica's. How does ignite know which two of the four directories to 
re-use? 

By default, folder name contains IP and port of node related to this
persistent store's data. So, the newly started node will choose a folder for
the IP and host that it has. Also, it's possible to set consistentId for the
node, in this case, the node will choose folder with the name related to
it's consistentId. 

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: How to use BinaryObject from existing data

[ezhuravlev]

Hi, 

In your case, you, most possibly use Id as key, and it could be Long or Int.
So, you just need to get it by this key. The example you provided describes
the case with complex primary key.

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Ignite 2.x upgrade guidelines

[ezhuravlev]

Hi,

At first, you may take a look at Ignite 2.0 migration guide:
https://cwiki.apache.org/confluence/display/IGNITE/Apache+Ignite+2.0+Migration+Guide

Also, the main difference in configuration between these versions is related
to the memory configuration, so take a look at this section from
documentation: https://apacheignite.readme.io/docs/memory-configuration

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: "HandshakeException: Remote node ID is not as expected" occurs when tested with 1000 clients

[ezhuravlev]

Hi,

What happened here: other node tried to connect to the node via address
cc3a7x2183.nhnsystem.com/10.112.96.139:5 and were not succeeded. After
it node tried to connect via other address, that was resolved by a node at a
start - 127.0.0.1:5. Of course, as these 2 nodes on the separate
machines, a node with different ID was found bound to this address.

To avoid communication via this localhost address, you can just specify a
host for communication Spi, so only this host will be used for communication
and not all resolved local addresses.

Also, the root cause here is not the wrong remote node ID, buy a problem
with communication to the node via first
address(cc3a7x2183.nhnsystem.com/10.112.96.139:5) it could be network
problems, firewall, maybe some GC pauses. So, you need to determine why a
connection was refused via normal address.

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: read/persist a huge cache by its affinity key

[ezhuravlev]

Hi Marco,

Why don't you want to use an Ignite persistence:
https://apacheignite.readme.io/docs/distributed-persistent-store or at least
Cache Store for a 3rd party DB:
https://apacheignite.readme.io/docs/3rd-party-store ?

By using Ignite persistence you will get everything you want here without
any additional changes.

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: copyOnRead to false

[ezhuravlev]

Hi Colin,

I would suggest you send a new message to user list with a detailed
description of your use case, I think the community could check it and give
some comments.

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Using event to reconnect spring created cache client to server

[ezhuravlev]

gunman524, I didn't say anything like that. 

I said that if you want to access cache from CQ, you need to start a new
thread

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Ignite 2.3.0 hangs in startup

[ezhuravlev]

Looks like you witnessed this issue:
https://issues.apache.org/jira/browse/IGNITE-6555
Which is already fixed and will be available in 2.4 release

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Using event to reconnect spring created cache client to server

[ezhuravlev]

It's not recommended to access caches from listeners. It's a general practice
to create a new thread after you get an event.

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Using event to reconnect spring created cache client to server

[ezhuravlev]

Try to start a new thread and access cache from it.

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Ignite 2.3.0 hangs in startup

[ezhuravlev]

For using slf4j, you need to have a configure file, as it showed in the
message, default it's config/java.util.logging.properties, so, you can add
this file to this path or try to configure slf4j:
https://www.slf4j.org/docs.html

Ignite slf4 logger constructor also have another constructor - new
Slf4jLogger(Logger logger).

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Grid freezing

[ezhuravlev]

Is it possible that in dequeuePortionIds you will have the same keys(at least
one the same key) for scanCache in different threads? If it's possible, I
think it's pretty easy to face a deadlock here with 
_getCurrentMax(customerMaxCache, queryScan), while I suppose 2 different
Scans could have the same customer(well, at least it could be deduced from
your comments).

Could you share full reproducer with the community? It's hard to try to find
a deadlock if you don't even have a source code.

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Failed to parse query: exception with Scala

[ezhuravlev]

Do you have indexed types for cache "sharedRDDProducer"?

like

cacheCfg.setIndexedTypes(String.class, String.class);

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Ignite 2.3.0 hangs in startup

[ezhuravlev]

Hi,

As I see in this error message:

 Failed to resolve default logging config file:
config/java.util.logging.properties Console logging handler is not
configured. 

You don't have java.util.logging.properties config file for SLF4j. I've
checked 2.3.0 binaries and this file exists in folder config. You need to
add it to your project.

Also, please check the documentation for configuring SLF4j logger, maybe you
missed something: 

https://apacheignite.readme.io/docs/logging#section-slf4j

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: How to configure a QueryEntity for a BinaryObject

[ezhuravlev]

Check QueryEntity class, it contains tableName property.

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Cache#invokeAsync cause StackOverflowError

[ezhuravlev]

Hi,

Create static inner class for EntryProcessor instead of lambda, it looks
like when lambda contains objects from outer class it could lead to
stackoverflow at the serialization.

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: REST API qryexe saying Argument cannot be null: sql.

[ezhuravlev]

Hi,

I don't see in your query defined cacheName and type parameters. If cache
name not provided, default cache will be used.

You can find all this information in the documentation:
https://apacheignite.readme.io/docs/rest-api#section-sql-query-execute

Also, have you checked logs from Ignite node?

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: client thread dumps

[ezhuravlev]

Hi,

Here is a description of all thread pools:
https://apacheignite.readme.io/v2.1/docs/thread-pools
You can try to reduce their size, but do it carefully, these changes could
lead to performance degradation.

If you will never use visor or rest-api you can also set
ConnectorConfiguration to null:

IgniteConfiguration cfg = new IgniteConfiguration(); 
cfg.setConnectorConfiguration(null); // disables the tcp rest connector 


Also, I found thread with pretty the same questions that you have about
threads:
http://apache-ignite-users.70518.x6.nabble.com/Ignite-Thread-count-td7636.html

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: copyOnRead to false

[ezhuravlev]

Hi,

Could you describe your benchmarks? So we can understand whats could be
wrong here.

If the value is fetched from a remote node, you will always get a copy. If
you get the value locally, you can force Ignite to return the stored
instance by setting CacheConfiguration.setCopyOnRead(false) property, but
this should be used only in read-only scenario. It's not safe to modify this
instance because the serialized form will not be updated until you call
cache.put(), so the one that reads it will potentially get the old value.
Additionally, it can be concurrently serialized which can cause data
corruption. 

If you use BinaryMarshaller, with copyOnRead flag, Ignite stores serialized
copy in heap and do not copy object on each "get" (if the object does not
change after "get"):

Please share info about your benchmarks(or code) to investigation of this
problem.

Evgenii 



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: About Apache Ignite Partitioned Cache

[ezhuravlev]

>The backups which is mentioned in the documentation, how do I define which
node is the primary node and >which node is the backup node.

It defined by Affinity Function, you can read about it here:
https://apacheignite.readme.io/docs/affinity-collocation#section-affinity-function

>I will have a four node cluster in two data centers , is it normal to think
each data center would have one >primary node and one backup node. What
would happen in the scenario that the primary node on one dc >cannot reach
the primary node on the other dc?

I think that you not fully understand how partitioned caches work. Node can
be primary for the part of the partitions, while other nodes will have other
parts as primary. Here is basic information about Partitioned cache that
should be enough to start:
https://apacheignite.readme.io/docs/cache-modes#section-partitioned-mode

>I assume that org.apache.ignite.spi.discovery.tcp.TcpDiscoverySpi is to be
used for the dsicoverySpi property >even for Partitioned Caches. Are there
special properties that should be filled in for the partitioned cache
>except for addresses and ports?

discoverySpi doesn't affect caches at all, so, you don't need to change
anything at org.apache.ignite.spi.discovery.tcp.TcpDiscoverySpi
configuration

>Most of the caches I am using is using uuid as keys , in that case how
would affinity collocation of the keys >work? Unfortunately I cannot change
the key structure in a short notice.

Do you want to collocate Data with Data or Compute with Data? In both cases
you can find information on this page:
https://apacheignite.readme.io/docs/affinity-collocation


Evgenii




--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Task management - MapReduce & ForkJoin performance penalty

[ezhuravlev]

Hi,

I've added Thread.sleep(200) to Jobs to simulate a small load.
Here is what I've got: 
1node: 1 Task 2000 Jobs ~25 sec
2nodes(on the same machine): 1 Task 2000 Jobs ~13 sec

What I want to say here - this overhead will be not noticeable on real Jobs.

What about some configuration changes - you could change
igniteConfiguration.setPublicThreadPoolSize(). By default, it set to CPU
count(with hyperthreading), but for small tasks you can make it bigger, for
example, CPU count*2.

Also, Ignite contains module benchmarks - you can check it and find
benchmarks for ComputeGrid.

All the best,
Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Serialization exception in Ignite compute job

[ezhuravlev]

Hi,

Non-static nested classes that implement Serializable must be defined in an
enclosing class that is also serializable. Non-static nested classes retain
an implicit reference to an instance of their enclosing class. If the
enclosing class is not serializable, the Java serialization mechanism fails
with a java.io.NotSerializableException. 

It's java rules, you can read about it here:
https://docs.oracle.com/javase/tutorial/java/javaOO/nested.html

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Possible starvation in striped pool. message

[ezhuravlev]

peerClassLoading used only for compute, for example for sharing job classes
between nodes, it's not working for objects that put into cache. If you want
to work without this classes on nodes, take a look to BinaryObjects:
https://apacheignite.readme.io/v2.0/docs/binary-marshaller

Evgenii



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Limiting the size of Persistent Store and clearing data on restart

[ezhuravlev]

Hi,

1) There is no possibility to limit Persistent Store size right now, but you
always can use separate disk volume with size 300GB for your case to use it
for Persistent Store.

2) Well it's not usual to use case, but after stopping node you can clear
data from work folder or just invoke removeAll on cache

Evgenii



--
View this message in context: 
http://apache-ignite-users.70518.x6.nabble.com/Limiting-the-size-of-Persistent-Store-and-clearing-data-on-restart-tp16507p16515.html
Sent from the Apache Ignite Users mailing list archive at Nabble.com.

Re: Ignite 2.0 transaction timeout not holding up

[ezhuravlev]

Hi,

This timeout is for the whole transaction, but when for the transaction you
usually need to take locks on different nodes. So, when the transaction was
succeeded in acquiring locks on one node, it sends remaining timeout o next
node, to acquire locks on it. 

For example, when you see timeout=59980, it means, most probably, that
transaction acquired locks on node1 in 20ms and sent remaining
timeout(59980ms) to node2 and on node2 it didn't succeed in acquiring locks.

Evgenii



--
View this message in context: 
http://apache-ignite-users.70518.x6.nabble.com/Ignite-2-0-transaction-timeout-not-holding-up-tp16462p16506.html
Sent from the Apache Ignite Users mailing list archive at Nabble.com.

Re: Error during activation of cluster

[ezhuravlev]

Hi,

To investigate this issue we need some additional information from your
nodes:

1. Persistence config from all nodes.
2. Catalog structure from C:\Data\IPD\WAL\ and from pathes from other nodes.
3. Checkpoint files from work/db/{NODE_ID}/cp - they are have a small size.

Evgenii



--
View this message in context: 
http://apache-ignite-users.70518.x6.nabble.com/Error-during-activation-of-cluster-tp16478p16499.html
Sent from the Apache Ignite Users mailing list archive at Nabble.com.

Re: Failed to wait for partition map exchange

[ezhuravlev]

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.

Which Java versions do you use in both cases? In eclipse and when you run it
with ignite.sh. 
I've seen problems with "Attempted to release write lock while not holding
it", but all of them were resolved after upgrading java version.

Evgenii



--
View this message in context: 
http://apache-ignite-users.70518.x6.nabble.com/Failed-to-wait-for-partition-map-exchange-tp16436p16495.html
Sent from the Apache Ignite Users mailing list archive at Nabble.com.

Re: Ignite Server and Client Configuration for different use cases

[ezhuravlev]

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.

You can find a lot of example for .Net here
https://github.com/apache/ignite/tree/master/modules/platforms/dotnet/examples

They cover most basic use cases, so you can use it for start to work with
Ignite. If you will need some more complicated, you can build your
configuration based on this examples or ask community to help in certain use
case.

Evgenii



--
View this message in context: 
http://apache-ignite-users.70518.x6.nabble.com/Ignite-Server-and-Client-Configuration-for-different-use-cases-tp16472p16479.html
Sent from the Apache Ignite Users mailing list archive at Nabble.com.

Re: Kerberos installation

[ezhuravlev]

Have you checked this link:
https://apacheignite-fs.readme.io/docs/file-system ?

You need to configure IGFS for hadoop and Ignite should use Hadoop api to
work with Kerberos

Evgenii



--
View this message in context: 
http://apache-ignite-users.70518.x6.nabble.com/Kerberos-installation-tp16200p16204.html
Sent from the Apache Ignite Users mailing list archive at Nabble.com.

Re: Activating Cluster taking too long

[ezhuravlev]

Hi,

Please share full logs from all nodes so I can help in investigating of your
problem.

Evgenii



--
View this message in context: 
http://apache-ignite-users.70518.x6.nabble.com/Activating-Cluster-taking-too-long-tp16093p16099.html
Sent from the Apache Ignite Users mailing list archive at Nabble.com.

Re: Caused by: org.h2.jdbc.JdbcSQLException: General error: "java.lang.IllegalMonitorStateException: Attempted to release write lock while not holding it

[ezhuravlev]

Hi,
I tried to run your code on Linux, but everything works fine for me. One
thing - I don't see SSL configuration in your config files, while your log
contains information that SSL was enabled. Is the same behavior reproducible
without SSL?

Also, I see that you use both ipv4 and ipv6, possibly it could lead to the
problem that was shown in the log - when a client was unable to connect to
the server at the first time. Try to run all nodes with
-Djava.net.preferIPv4Stack=true property.






--
View this message in context: 
http://apache-ignite-users.70518.x6.nabble.com/Caused-by-org-h2-jdbc-JdbcSQLException-General-error-java-lang-IllegalMonitorStateException-Attemptet-tp15684p16046.html
Sent from the Apache Ignite Users mailing list archive at Nabble.com.

Re: Ignite with Camel

[ezhuravlev]

Have you checked this page: http://camel.apache.org/ignite.html ?



--
View this message in context: 
http://apache-ignite-users.70518.x6.nabble.com/Ignite-with-Camel-tp13687p16030.html
Sent from the Apache Ignite Users mailing list archive at Nabble.com.

Re: Remote node ID is not as expected - New Node not coming up

[ezhuravlev]

Hello,

Looks like you have you have nodes on different hosts with same ports. Do
you have other exceptions in log? Even if he failed to connect to this node
through localhost address, it should try to connect to other addresses of
node fc2bcbe5-2a4b-406b-b60d-69a445f52fd9. Also, on which addreses this node
were binded?

To avoid it you can set different locaPort on different nodes, even if they
hosts on different machines or set manually localHost for each node.

Evgenii



--
View this message in context: 
http://apache-ignite-users.70518.x6.nabble.com/Remote-node-ID-is-not-as-expected-New-Node-not-coming-up-tp15858p15940.html
Sent from the Apache Ignite Users mailing list archive at Nabble.com.

Re: Can not deserialize ignite cache expired event's oldValue

[ezhuravlev]

Hi,

You've got instance of BinaryObject, not instance of
com.coupang.ignite.fds.model.DiscountCouponOrderItem.

To get instance of this class you need to deserialise BinaryObject value by
invoking oldVal.deserialize()

Evgenii



--
View this message in context: 
http://apache-ignite-users.70518.x6.nabble.com/Can-not-deserialize-ignite-cache-expired-event-s-oldValue-tp15848p15870.html
Sent from the Apache Ignite Users mailing list archive at Nabble.com.

Re: Cache Store in Ignite2.0

[ezhuravlev]

Hi,

If you want, you can work without POJO at all. It's possible to work wih
BinaryObjects: https://apacheignite.readme.io/docs/binary-marshaller

Here is working example:
http://apache-ignite-users.70518.x6.nabble.com/SQL-queries-with-BinaryObject-td5636.html

Evgenii



--
View this message in context: 
http://apache-ignite-users.70518.x6.nabble.com/Cache-Store-in-Ignite2-0-tp15227p15230.html
Sent from the Apache Ignite Users mailing list archive at Nabble.com.

Re: Data loss validation. TopologyValidator

[ezhuravlev]

Hi,

Yes, TopologyValidator was implemented exactly for these things.

Also, it's possible to listen for EVT_CACHE_REBALANCE_PART_DATA_LOST events
and implement some logic on it, but I think that TopologyValidator is better
solution.

Evgenii



--
View this message in context: 
http://apache-ignite-users.70518.x6.nabble.com/Data-loss-validation-TopologyValidator-tp15147p15172.html
Sent from the Apache Ignite Users mailing list archive at Nabble.com.

Re: How does node vistor task can work?

[ezhuravlev]

You don't need to use Visor tasks to get statistics.

Try this:

ignite.cluster().metrics()
or 
ignite.cluster().node(node).metrics()

Evgeniii



--
View this message in context: 
http://apache-ignite-users.70518.x6.nabble.com/How-does-node-vistor-task-can-work-tp15095p15121.html
Sent from the Apache Ignite Users mailing list archive at Nabble.com.

Re: Provide persistence settings inline for Cassandra KeyValuePersistenceSettings?

[ezhuravlev]

Hi,

It's possible to provide settings with String:








Evgenii



--
View this message in context: 
http://apache-ignite-users.70518.x6.nabble.com/Provide-persistence-settings-inline-for-Cassandra-KeyValuePersistenceSettings-tp14724p14802.html
Sent from the Apache Ignite Users mailing list archive at Nabble.com.

Re: Ignite not conneting to remote node if Using FQDN

[ezhuravlev]

Hi,

Could you provide your config files and logs?

Evgenii



--
View this message in context: 
http://apache-ignite-users.70518.x6.nabble.com/Ignite-not-conneting-to-remote-node-if-Using-FQDN-tp14634p14801.html
Sent from the Apache Ignite Users mailing list archive at Nabble.com.

Re: ignite can not find other ignite clients with attched config xml

[ezhuravlev]

Hi,

Please provide logs from Ignite nodes, so we could investigate them.

Evgenii



--
View this message in context: 
http://apache-ignite-users.70518.x6.nabble.com/ignite-can-not-find-other-ignite-clients-with-attched-config-xml-tp14613p14749.html
Sent from the Apache Ignite Users mailing list archive at Nabble.com.

Re: Cache creation before making any operation the cache?

[ezhuravlev]

Hi,

On user operations use Ignite.cache instead of Ignite.getOrCreateCache. it
will throw an exception if cache doesn't exist.
Or, you can get list of cache names by Ignite.cacheNames().

Evgenii



--
View this message in context: 
http://apache-ignite-users.70518.x6.nabble.com/Cache-creation-before-making-any-operation-the-cache-tp14264p14267.html
Sent from the Apache Ignite Users mailing list archive at Nabble.com.

Re: Putting a multi-level object into Ignite cache hangs on put()

[ezhuravlev]

It works for me without hangs, here is log: 

Created Ignite cache: IgniteCacheProxy [delegate=GridDhtAtomicCache
[defRes=org.apache.ignite.internal.processors.cache.distributed.dht.atomic.GridDhtAtomicCache$1@72f46e16,
near=null, super=GridDhtCacheAdapter
[multiTxHolder=java.lang.ThreadLocal@71104a4,
super=GridDistributedCacheAdapter [super=GridCacheAdapter
[locMxBean=org.apache.ignite.internal.processors.cache.CacheLocalMetricsMXBeanImpl@664a9613,
clusterMxBean=org.apache.ignite.internal.processors.cache.CacheClusterMetricsMXBeanImpl@5118388b,
aff=org.apache.ignite.internal.processors.cache.affinity.GridCacheAffinityImpl@15a902e7,
igfsDataCache=false, mongoDataCache=false, mongoMetaCache=false,
igfsDataCacheSize=null,
asyncOpsSem=java.util.concurrent.Semaphore@7876d598[Permits = 500],
name=myInputFields1Cache, size=0, opCtx=null]
adding id=0
adding id=1
adding id=2
adding id=3
adding id=4
adding id=5
adding id=6
adding id=7
adding id=8
adding id=9
results in cache:10
Got val=com.rentd.service.cache.vo.InputFields1@2c95ac9e
Got val=com.rentd.service.cache.vo.InputFields1@4e4efc1b
Got val=com.rentd.service.cache.vo.InputFields1@459f7aa3
Got val=com.rentd.service.cache.vo.InputFields1@7cc586a8
Got val=com.rentd.service.cache.vo.InputFields1@7db534f2
Got val=com.rentd.service.cache.vo.InputFields1@44a2b17b
Got val=com.rentd.service.cache.vo.InputFields1@7a56812e
Got val=com.rentd.service.cache.vo.InputFields1@2a76b80a
Got val=com.rentd.service.cache.vo.InputFields1@7eb01b12
Got val=com.rentd.service.cache.vo.InputFields1@2f4854d6
[11:08:32] Ignite node stopped OK [name=RentdGrid, uptime=00:00:00:363]

One thing - I've run it from Idea, not Eclipse, but it's shouldn't affect
this. So, could you share your logs?

Evgenii



--
View this message in context: 
http://apache-ignite-users.70518.x6.nabble.com/Putting-a-multi-level-object-into-Ignite-cache-hangs-on-put-tp14086p14158.html
Sent from the Apache Ignite Users mailing list archive at Nabble.com.

Re: It seems WebSession's removeAttribute does not support HttpSessionBindingListener

[ezhuravlev]

duplicate for
http://apache-ignite-developers.2346864.n4.nabble.com/It-seems-WebSession-s-removeAttribute-does-not-support-HttpSessionBindingListener-td19184.html



--
View this message in context: 
http://apache-ignite-users.70518.x6.nabble.com/It-seems-WebSession-s-removeAttribute-does-not-support-HttpSessionBindingListener-tp14099p14157.html
Sent from the Apache Ignite Users mailing list archive at Nabble.com.

Re: Ignite Custom Java Code calling

[ezhuravlev]

Hi,

It's not clear what do you want to do, but javadoc for QuerySqlFunction
contains some sample code for usage it:
https://ignite.apache.org/releases/mobile/org/apache/ignite/cache/query/annotations/QuerySqlFunction.html

How do you want to use HA with cluster singleton?

What do you mean by Ignite native client? Client node?

Evgenii



--
View this message in context: 
http://apache-ignite-users.70518.x6.nabble.com/Ignite-Custom-Java-Code-calling-tp13986p14156.html
Sent from the Apache Ignite Users mailing list archive at Nabble.com.

Re: ignite compute debugging

[ezhuravlev]

Hi,

In this message after some information also printed a full stacktrace for
this exception. Just check your log closely. 

If you don't have this stacktrace, please, share full logs with us.

Evgenii



--
View this message in context: 
http://apache-ignite-users.70518.x6.nabble.com/ignite-compute-debugging-tp14123p14141.html
Sent from the Apache Ignite Users mailing list archive at Nabble.com.

Re: Camel-Ignite cache query : Could not find a suitable setter for property: query

[ezhuravlev]

Hi,

You're trying to put in option "query", which have a type Query, a value,
which have a type String.
Do you have a registry bean #query1?

This documentation has a descriptions for parameters:
http://camel.apache.org/ignite.html

Evgenii



--
View this message in context: 
http://apache-ignite-users.70518.x6.nabble.com/Camel-Ignite-cache-query-Could-not-find-a-suitable-setter-for-property-query-tp14128p14138.html
Sent from the Apache Ignite Users mailing list archive at Nabble.com.

Re: Ignite spring transactions integration works?

[ezhuravlev]

Hi,

Have you enabled spring transaction support, as it described here?

https://ignite.apache.org/releases/2.0.0/javadoc/org/apache/ignite/transactions/spring/SpringTransactionManager.html

Evgenii



--
View this message in context: 
http://apache-ignite-users.70518.x6.nabble.com/Ignite-spring-transactions-integration-works-tp14050p14135.html
Sent from the Apache Ignite Users mailing list archive at Nabble.com.

Re: Putting a multi-level object into Ignite cache hangs on put()

[ezhuravlev]

Ok, then send full reproducer as a maven project, I will try it with spring.
I don't have AppConfig.class.

Evgenii



--
View this message in context: 
http://apache-ignite-users.70518.x6.nabble.com/Putting-a-multi-level-object-into-Ignite-cache-hangs-on-put-tp14086p14134.html
Sent from the Apache Ignite Users mailing list archive at Nabble.com.

Re: Failed to register query type: QueryTypeDescriptorImpl

[ezhuravlev]

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.

I tried your code, but it's works without problems for me. One thing - I
changed your InputFields class to my TestClass, so it could be that this
class is a root of a problem. Could you share it with us?

Evgenii



--
View this message in context: 
http://apache-ignite-users.70518.x6.nabble.com/Failed-to-register-query-type-QueryTypeDescriptorImpl-tp14052p14133.html
Sent from the Apache Ignite Users mailing list archive at Nabble.com.

Re: Putting a multi-level object into Ignite cache hangs on put()

[ezhuravlev]

Hi, 

I've tried to put same values to local ignite node without client and
spring, it works without any problems.

Could you share logs from both server and client node?

Evgenii



--
View this message in context: 
http://apache-ignite-users.70518.x6.nabble.com/Putting-a-multi-level-object-into-Ignite-cache-hangs-on-put-tp14086p14131.html
Sent from the Apache Ignite Users mailing list archive at Nabble.com.

Re: [TCP Discovery] connection ip configuration

[ezhuravlev]

Hi,

You can specify localHost in IgniteConfiguration:
https://ignite.apache.org/releases/2.0.0/javadoc/org/apache/ignite/configuration/IgniteConfiguration.html#setLocalHost(java.lang.String).
For example:



 

Evgenii



--
View this message in context: 
http://apache-ignite-users.70518.x6.nabble.com/TCP-Discovery-connection-ip-configuration-tp14075p14076.html
Sent from the Apache Ignite Users mailing list archive at Nabble.com.