Number of replicas


(Marcin Dojwa) #1

Hi,

I have the following situation:

  • 3 nodes
  • 1 index with 60 shards and 2 replicas
  • NO data yet

After starting all nodes I have all 180 shards as 'unassigned' in
/_cluster/health. After changing the index setting to have only 1 replica
all the shards got activated.
My question is: what is the reason of this and is there any connection
between number of replicas, number of shards and minimum number of nodes ?

Thank you for help.
Best regards.
Marcin Dojwa

--


(Marcin Dojwa) #2

I have default settings for all nodes, all nodes are the same.

2012/10/24 Marcin Dojwa m.dojwa@livechatinc.com

Hi,

I have the following situation:

  • 3 nodes
  • 1 index with 60 shards and 2 replicas
  • NO data yet

After starting all nodes I have all 180 shards as 'unassigned' in
/_cluster/health. After changing the index setting to have only 1 replica
all the shards got activated.
My question is: what is the reason of this and is there any connection
between number of replicas, number of shards and minimum number of nodes ?

Thank you for help.
Best regards.
Marcin Dojwa

--


(Igor Motov) #3

Interesting. What happens if you change number of replicas back to 2? Can
you recreate the problem and post complete output of the /_cluster/health
command?

On Wednesday, October 24, 2012 1:57:41 AM UTC-4, Marcin Dojwa wrote:

I have default settings for all nodes, all nodes are the same.

2012/10/24 Marcin Dojwa <m.d...@livechatinc.com <javascript:>>

Hi,

I have the following situation:

  • 3 nodes
  • 1 index with 60 shards and 2 replicas
  • NO data yet

After starting all nodes I have all 180 shards as 'unassigned' in
/_cluster/health. After changing the index setting to have only 1 replica
all the shards got activated.
My question is: what is the reason of this and is there any connection
between number of replicas, number of shards and minimum number of nodes ?

Thank you for help.
Best regards.
Marcin Dojwa

--


(Marcin Dojwa) #4

Hi,

After setting it back to 2 replicas all shards gets assigned. It works fine
now even if I stop all nodes and start them again. The problem was when I
started new cluster with 3 nodes, 1 index (60 shards and 2 replicas) and NO
data. All the shards got assigned after changing replicas to 1. Now it
looks like it works, but I suggest checking this by the author if this is
not a bigger problem.

This is how it looked like at the beginning.

  1. There are 3 nodes with 1 index with 60 shards and 2 replicas (all nodes
    are stopped)

  2. After starting 1st node /_cluster/health:
    {

    • cluster_name: "production",
    • status: "red",
    • timed_out: false,
    • number_of_nodes: 1,
    • number_of_data_nodes: 1,
    • active_primary_shards: 0,
    • active_shards: 0,
    • relocating_shards: 0,
    • initializing_shards: 0,
    • unassigned_shards: 180

}

  1. After starting 2nd node:
    {

    • cluster_name: "production",
    • status: "red",
    • timed_out: false,
    • number_of_nodes: 2,
    • number_of_data_nodes: 2,
    • active_primary_shards: 0,
    • active_shards: 0,
    • relocating_shards: 0,
    • initializing_shards: 0,
    • unassigned_shards: 180

}

  1. After starting 3rd node:
    {

    • cluster_name: "production",
    • status: "red",
    • timed_out: false,
    • number_of_nodes: 3,
    • number_of_data_nodes: 3,
    • active_primary_shards: 0,
    • active_shards: 0,
    • relocating_shards: 0,
    • initializing_shards: 0,
    • unassigned_shards: 180

}

  1. After changing replicas to 1:
    {

    • cluster_name: "production",
    • status: "red",
    • timed_out: false,
    • number_of_nodes: 3,
    • number_of_data_nodes: 3,
    • active_primary_shards: 60,
    • active_shards: 120,
    • relocating_shards: 0,
    • initializing_shards: 0,
    • unassigned_shards: 0

}

Best regards.

2012/10/24 Igor Motov imotov@gmail.com

Interesting. What happens if you change number of replicas back to 2? Can
you recreate the problem and post complete output of the /_cluster/health
command?

On Wednesday, October 24, 2012 1:57:41 AM UTC-4, Marcin Dojwa wrote:

I have default settings for all nodes, all nodes are the same.

2012/10/24 Marcin Dojwa m.d...@livechatinc.com

Hi,

I have the following situation:

  • 3 nodes
  • 1 index with 60 shards and 2 replicas
  • NO data yet

After starting all nodes I have all 180 shards as 'unassigned' in
/_cluster/health. After changing the index setting to have only 1 replica
all the shards got activated.
My question is: what is the reason of this and is there any connection
between number of replicas, number of shards and minimum number of nodes ?

Thank you for help.
Best regards.
Marcin Dojwa

--

--


(Ivan Brusic) #5

Did you change discovery.zen.minimum_master_nodes? Probably would not
account for this behavior, just wondering what your value is.

On Thu, Oct 25, 2012 at 12:56 AM, Marcin Dojwa m.dojwa@livechatinc.comwrote:

Hi,

After setting it back to 2 replicas all shards gets assigned. It works
fine now even if I stop all nodes and start them again. The problem was
when I started new cluster with 3 nodes, 1 index (60 shards and 2 replicas)
and NO data. All the shards got assigned after changing replicas to 1. Now
it looks like it works, but I suggest checking this by the author if this
is not a bigger problem.

This is how it looked like at the beginning.

  1. There are 3 nodes with 1 index with 60 shards and 2 replicas (all nodes
    are stopped)

  2. After starting 1st node /_cluster/health:
    {

    • cluster_name: "production",
    • status: "red",
    • timed_out: false,
    • number_of_nodes: 1,
    • number_of_data_nodes: 1,
    • active_primary_shards: 0,
    • active_shards: 0,
    • relocating_shards: 0,
    • initializing_shards: 0,
    • unassigned_shards: 180

}

  1. After starting 2nd node:
    {

    • cluster_name: "production",
    • status: "red",
    • timed_out: false,
    • number_of_nodes: 2,
    • number_of_data_nodes: 2,
    • active_primary_shards: 0,
    • active_shards: 0,
    • relocating_shards: 0,
    • initializing_shards: 0,
    • unassigned_shards: 180

}

  1. After starting 3rd node:
    {

    • cluster_name: "production",
    • status: "red",
    • timed_out: false,
    • number_of_nodes: 3,
    • number_of_data_nodes: 3,
    • active_primary_shards: 0,
    • active_shards: 0,
    • relocating_shards: 0,
    • initializing_shards: 0,
    • unassigned_shards: 180

}

  1. After changing replicas to 1:
    {

    • cluster_name: "production",
    • status: "red",
    • timed_out: false,
    • number_of_nodes: 3,
    • number_of_data_nodes: 3,
    • active_primary_shards: 60,
    • active_shards: 120,
    • relocating_shards: 0,
    • initializing_shards: 0,
    • unassigned_shards: 0

}

Best regards.

2012/10/24 Igor Motov imotov@gmail.com

Interesting. What happens if you change number of replicas back to 2? Can
you recreate the problem and post complete output of the /_cluster/health
command?

On Wednesday, October 24, 2012 1:57:41 AM UTC-4, Marcin Dojwa wrote:

I have default settings for all nodes, all nodes are the same.

2012/10/24 Marcin Dojwa m.d...@livechatinc.com

Hi,

I have the following situation:

  • 3 nodes
  • 1 index with 60 shards and 2 replicas
  • NO data yet

After starting all nodes I have all 180 shards as 'unassigned' in
/_cluster/health. After changing the index setting to have only 1 replica
all the shards got activated.
My question is: what is the reason of this and is there any connection
between number of replicas, number of shards and minimum number of nodes ?

Thank you for help.
Best regards.
Marcin Dojwa

--

--

--


(Marcin Dojwa) #6

Hi, I didn't change this value, it is set to default.

W dniu czwartek, 25 października 2012 użytkownik Ivan Brusic napisał:

Did you change discovery.zen.minimum_master_nodes? Probably would not
account for this behavior, just wondering what your value is.

On Thu, Oct 25, 2012 at 12:56 AM, Marcin Dojwa <m.dojwa@livechatinc.com<javascript:_e({}, 'cvml', 'm.dojwa@livechatinc.com');>

wrote:

Hi,

After setting it back to 2 replicas all shards gets assigned. It works
fine now even if I stop all nodes and start them again. The problem was
when I started new cluster with 3 nodes, 1 index (60 shards and 2 replicas)
and NO data. All the shards got assigned after changing replicas to 1. Now
it looks like it works, but I suggest checking this by the author if this
is not a bigger problem.

This is how it looked like at the beginning.

  1. There are 3 nodes with 1 index with 60 shards and 2 replicas (all
    nodes are stopped)

  2. After starting 1st node /_cluster/health:
    {

    • cluster_name: "production",
    • status: "red",
    • timed_out: false,
    • number_of_nodes: 1,
    • number_of_data_nodes: 1,
    • active_primary_shards: 0,
    • active_shards: 0,
    • relocating_shards: 0,
    • initializing_shards: 0,
    • unassigned_shards: 180

}

  1. After starting 2nd node:
    {

    • cluster_name: "production",
    • status: "red",
    • timed_out: false,

--

--

--


(system) #7