Next steps... how to reboot a node without starting a re-balance of unallocated shards


(Aaron Anderson) #1

I have a 4 node cluster. (just a proof of concept right now, but will
potentially consume 100gb+ per day)

What configuration items do I need to learn/read about to delay
re-balancing of unallocated shards? For example, I am required to do system
patching on a regular basis, so each node is going to have to go through a
reboot. How can I reboot the host, or restart ES for new settings to take
effect, without kicking off the re-balance?

Can I somehow specify a delay? My current config is posted below.

ALSO: I notice (in the head plugin) if I do a "service elasticsearch
restart" on a node that I will end up with two nodes with the same address
each holding different shards, probably because in the short time the node
was down that particular shard wasn't moved somewhere else. Is this OK? For
example, if i restart LouDvEs001 (which is the address, names are auto gen)
I'll have two nodes with different names, but the same address.

cluster:
name: dvlogstash
discovery:
zen:
minimum_master_nodes: 2
ping:
multicast:
enabled: false
unicast:
hosts:
- loudves001
- loudves002
- loudves003
- loudves004
path:
conf: /etc/elasticsearch/
logs: /var/log/elasticsearch/

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to elasticsearch+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/elasticsearch/8b7f59db-50fe-4df2-9f18-167d66fccab7%40googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.


(Mark Walkom) #2

cluster.routing.allocation.disable_allocation is probably what you want.

Regards,
Mark Walkom

Infrastructure Engineer
Campaign Monitor
email: markw@campaignmonitor.com
web: www.campaignmonitor.com

On 16 December 2013 11:31, Aaron Anderson anderson.aaron@gmail.com wrote:

I have a 4 node cluster. (just a proof of concept right now, but will
potentially consume 100gb+ per day)

What configuration items do I need to learn/read about to delay
re-balancing of unallocated shards? For example, I am required to do system
patching on a regular basis, so each node is going to have to go through a
reboot. How can I reboot the host, or restart ES for new settings to take
effect, without kicking off the re-balance?

Can I somehow specify a delay? My current config is posted below.

ALSO: I notice (in the head plugin) if I do a "service elasticsearch
restart" on a node that I will end up with two nodes with the same address
each holding different shards, probably because in the short time the node
was down that particular shard wasn't moved somewhere else. Is this OK? For
example, if i restart LouDvEs001 (which is the address, names are auto gen)
I'll have two nodes with different names, but the same address.

cluster:
name: dvlogstash
discovery:
zen:
minimum_master_nodes: 2
ping:
multicast:
enabled: false
unicast:
hosts:
- loudves001
- loudves002
- loudves003
- loudves004
path:
conf: /etc/elasticsearch/
logs: /var/log/elasticsearch/

--
You received this message because you are subscribed to the Google Groups
"elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an
email to elasticsearch+unsubscribe@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/elasticsearch/8b7f59db-50fe-4df2-9f18-167d66fccab7%40googlegroups.com
.
For more options, visit https://groups.google.com/groups/opt_out.

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to elasticsearch+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/elasticsearch/CAEM624bsBStCWZ%2BA6e13V3LbSNGztPbNGhPuQA26NeMKwMhDLQ%40mail.gmail.com.
For more options, visit https://groups.google.com/groups/opt_out.


(Ivan Brusic) #3

Besides what Mark mentioned (the key setting to use), the only other thing
I can suggest is to flush the transaction log before rebooting any node
(assuming you can disable indexing and allocation).

Cheers,

Ivan

On Sun, Dec 15, 2013 at 4:56 PM, Mark Walkom markw@campaignmonitor.comwrote:

cluster.routing.allocation.disable_allocation is probably what you want.

Regards,
Mark Walkom

Infrastructure Engineer
Campaign Monitor
email: markw@campaignmonitor.com
web: www.campaignmonitor.com

On 16 December 2013 11:31, Aaron Anderson anderson.aaron@gmail.comwrote:

I have a 4 node cluster. (just a proof of concept right now, but will
potentially consume 100gb+ per day)

What configuration items do I need to learn/read about to delay
re-balancing of unallocated shards? For example, I am required to do system
patching on a regular basis, so each node is going to have to go through a
reboot. How can I reboot the host, or restart ES for new settings to take
effect, without kicking off the re-balance?

Can I somehow specify a delay? My current config is posted below.

ALSO: I notice (in the head plugin) if I do a "service elasticsearch
restart" on a node that I will end up with two nodes with the same address
each holding different shards, probably because in the short time the node
was down that particular shard wasn't moved somewhere else. Is this OK? For
example, if i restart LouDvEs001 (which is the address, names are auto gen)
I'll have two nodes with different names, but the same address.

cluster:
name: dvlogstash
discovery:
zen:
minimum_master_nodes: 2
ping:
multicast:
enabled: false
unicast:
hosts:
- loudves001
- loudves002
- loudves003
- loudves004
path:
conf: /etc/elasticsearch/
logs: /var/log/elasticsearch/

--
You received this message because you are subscribed to the Google Groups
"elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an
email to elasticsearch+unsubscribe@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/elasticsearch/8b7f59db-50fe-4df2-9f18-167d66fccab7%40googlegroups.com
.
For more options, visit https://groups.google.com/groups/opt_out.

--
You received this message because you are subscribed to the Google Groups
"elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an
email to elasticsearch+unsubscribe@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/elasticsearch/CAEM624bsBStCWZ%2BA6e13V3LbSNGztPbNGhPuQA26NeMKwMhDLQ%40mail.gmail.com
.

For more options, visit https://groups.google.com/groups/opt_out.

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to elasticsearch+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/elasticsearch/CALY%3DcQCBRiRzfy70MkPHNZ5wFk9C3N%2BueH6C3UmOpX5j0tw-vg%40mail.gmail.com.
For more options, visit https://groups.google.com/groups/opt_out.


(system) #4