You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@ignite.apache.org by torjt <je...@gmail.com> on 2017/10/16 21:54:08 UTC
Client Near Cache Configuration Lost after Cluster Node Removed
Hello All,
We are having an issue with Ignite client near caches being "lost" upon a
cluster node being removed from the Ignite cluster. Furthermore, using
version 2.1.0, we are seeing this issue when another client joins the
topology. I built Ignite from GIT today, 10/16/17, with the latest changes,
ver. 2.3.0-SNAPSHOT. As of version 2.3.0-SNAPSHOT, bringing clients up/down
does not cause an active client to lose its near cache and performance is
good. However, when we remove a node from the cluster, the client
immediately communicates with the cluster and disregards its near cache.
Restarting the client remedies the issue.
The following is the steps to reproduce the issue:
Apache Ignite Version:
*SNAPSHOT#20171016-sha1:ca6662bcb4eecc62493e2e25a572ed0b982c046c*
1. Start 2 Ignite servers
2. Start client with caches configured as near-cache
3. Access caches
4. Stop node client is connected to
4a. Client immediately bypasses near cache and access "cluster" for cache
miss
--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/
Re: Client Near Cache Configuration Lost after Cluster Node Removed
Posted by Timay <to...@trustwave.com>.
I believe i had the same issue, i have posted a test and my finding to the
user group. Which can be found here.
http://apache-ignite-users.70518.x6.nabble.com/Near-Cache-Topoolgy-change-causes-NearCache-to-always-miss-td17539.html
--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/
Re: Client Near Cache Configuration Lost after Cluster Node Removed
Posted by torjt <je...@gmail.com>.
To answer your question, using Visor, one can see cache misses increase upon
the scenario described. Furthermore, measure cache.get() calls increase
from sub millisecond to 20 to 30 miillis.
--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/
Re: Client Near Cache Configuration Lost after Cluster Node Removed
Posted by Nikolai Tikhonov <nt...@apache.org>.
Hello,
Could you say how you determine that client node loads data from remote
node bypasses near cache?
I'm not able reproduce this behaviour locally, could you share a simple
maven project that reproduces this behaviour?
On Tue, Oct 17, 2017 at 12:54 AM, torjt <je...@gmail.com> wrote:
> Hello All,
>
> We are having an issue with Ignite client near caches being "lost" upon a
> cluster node being removed from the Ignite cluster. Furthermore, using
> version 2.1.0, we are seeing this issue when another client joins the
> topology. I built Ignite from GIT today, 10/16/17, with the latest
> changes,
> ver. 2.3.0-SNAPSHOT. As of version 2.3.0-SNAPSHOT, bringing clients
> up/down
> does not cause an active client to lose its near cache and performance is
> good. However, when we remove a node from the cluster, the client
> immediately communicates with the cluster and disregards its near cache.
> Restarting the client remedies the issue.
>
> The following is the steps to reproduce the issue:
> Apache Ignite Version:
> *SNAPSHOT#20171016-sha1:ca6662bcb4eecc62493e2e25a572ed0b982c046c*
> 1. Start 2 Ignite servers
> 2. Start client with caches configured as near-cache
> 3. Access caches
> 4. Stop node client is connected to
> 4a. Client immediately bypasses near cache and access "cluster" for cache
> miss
>
>
>
>
>
> --
> Sent from: http://apache-ignite-users.70518.x6.nabble.com/
>
Re: Client Near Cache Configuration Lost after Cluster Node Removed
Posted by Denis Magda <dm...@apache.org>.
+ user list
> On Oct 20, 2017, at 11:00 PM, Denis Magda <dm...@apache.org> wrote:
>
> Ignite caching experts please chime in.
>
>> Begin forwarded message:
>>
>> From: torjt <je...@gmail.com>
>> Subject: Client Near Cache Configuration Lost after Cluster Node Removed
>> Date: October 16, 2017 at 2:54:08 PM PDT
>> To: user@ignite.apache.org
>> Reply-To: user@ignite.apache.org
>>
>> Hello All,
>>
>> We are having an issue with Ignite client near caches being "lost" upon a
>> cluster node being removed from the Ignite cluster. Furthermore, using
>> version 2.1.0, we are seeing this issue when another client joins the
>> topology. I built Ignite from GIT today, 10/16/17, with the latest changes,
>> ver. 2.3.0-SNAPSHOT. As of version 2.3.0-SNAPSHOT, bringing clients up/down
>> does not cause an active client to lose its near cache and performance is
>> good. However, when we remove a node from the cluster, the client
>> immediately communicates with the cluster and disregards its near cache.
>> Restarting the client remedies the issue.
>>
>> The following is the steps to reproduce the issue:
>> Apache Ignite Version:
>> *SNAPSHOT#20171016-sha1:ca6662bcb4eecc62493e2e25a572ed0b982c046c*
>> 1. Start 2 Ignite servers
>> 2. Start client with caches configured as near-cache
>> 3. Access caches
>> 4. Stop node client is connected to
>> 4a. Client immediately bypasses near cache and access "cluster" for cache
>> miss
>>
>>
>>
>>
>>
>> --
>> Sent from: http://apache-ignite-users.70518.x6.nabble.com/
>
Re: Client Near Cache Configuration Lost after Cluster Node Removed
Posted by Denis Magda <dm...@apache.org>.
+ user list
> On Oct 20, 2017, at 11:00 PM, Denis Magda <dm...@apache.org> wrote:
>
> Ignite caching experts please chime in.
>
>> Begin forwarded message:
>>
>> From: torjt <je...@gmail.com>
>> Subject: Client Near Cache Configuration Lost after Cluster Node Removed
>> Date: October 16, 2017 at 2:54:08 PM PDT
>> To: user@ignite.apache.org
>> Reply-To: user@ignite.apache.org
>>
>> Hello All,
>>
>> We are having an issue with Ignite client near caches being "lost" upon a
>> cluster node being removed from the Ignite cluster. Furthermore, using
>> version 2.1.0, we are seeing this issue when another client joins the
>> topology. I built Ignite from GIT today, 10/16/17, with the latest changes,
>> ver. 2.3.0-SNAPSHOT. As of version 2.3.0-SNAPSHOT, bringing clients up/down
>> does not cause an active client to lose its near cache and performance is
>> good. However, when we remove a node from the cluster, the client
>> immediately communicates with the cluster and disregards its near cache.
>> Restarting the client remedies the issue.
>>
>> The following is the steps to reproduce the issue:
>> Apache Ignite Version:
>> *SNAPSHOT#20171016-sha1:ca6662bcb4eecc62493e2e25a572ed0b982c046c*
>> 1. Start 2 Ignite servers
>> 2. Start client with caches configured as near-cache
>> 3. Access caches
>> 4. Stop node client is connected to
>> 4a. Client immediately bypasses near cache and access "cluster" for cache
>> miss
>>
>>
>>
>>
>>
>> --
>> Sent from: http://apache-ignite-users.70518.x6.nabble.com/
>
Fwd: Client Near Cache Configuration Lost after Cluster Node Removed
Posted by Denis Magda <dm...@apache.org>.
Ignite caching experts please chime in.
> Begin forwarded message:
>
> From: torjt <je...@gmail.com>
> Subject: Client Near Cache Configuration Lost after Cluster Node Removed
> Date: October 16, 2017 at 2:54:08 PM PDT
> To: user@ignite.apache.org
> Reply-To: user@ignite.apache.org
>
> Hello All,
>
> We are having an issue with Ignite client near caches being "lost" upon a
> cluster node being removed from the Ignite cluster. Furthermore, using
> version 2.1.0, we are seeing this issue when another client joins the
> topology. I built Ignite from GIT today, 10/16/17, with the latest changes,
> ver. 2.3.0-SNAPSHOT. As of version 2.3.0-SNAPSHOT, bringing clients up/down
> does not cause an active client to lose its near cache and performance is
> good. However, when we remove a node from the cluster, the client
> immediately communicates with the cluster and disregards its near cache.
> Restarting the client remedies the issue.
>
> The following is the steps to reproduce the issue:
> Apache Ignite Version:
> *SNAPSHOT#20171016-sha1:ca6662bcb4eecc62493e2e25a572ed0b982c046c*
> 1. Start 2 Ignite servers
> 2. Start client with caches configured as near-cache
> 3. Access caches
> 4. Stop node client is connected to
> 4a. Client immediately bypasses near cache and access "cluster" for cache
> miss
>
>
>
>
>
> --
> Sent from: http://apache-ignite-users.70518.x6.nabble.com/