I have a question about the solution to solve the "Shard unassigned"
problem.
Here my setup, pretty simple:
1 NODE - 2 SHARDS / index - NO REPLICA
The situation is the following:
I manually moved all the content of /var/data/elasticsearch to another
location for backup need;
So I am able to start with a fresh cluster.
I restarted the service and everything was going fine until I realised
that all my Kibana dashboards were not available on the newly live
cluster.
I understood it was because of the previous move of the files, so I
simply moved the* kibana-int* index directory from the backup dir back to /var/data/elasticsearch
If it was so simple...
The new cluster showed the old kibana-int index as unassigned despite
restart of the service.
Questions:
How the new cluster can unassigned this index (does it see that it does
not created this index itself based on timestamp or other, and therefore
decide to not take it...) ?
How can I assign this index ? The only solution I know is to start
another node (so a 2 nodes cluster) hoping that the second node will take
the* kibana-int *and then move this index onto the first node.
Thanks a lot for sharing your experience on this matters, I am sure it will
help other.
It's probably because the cluster state doesn't know about the index even
though it can see it on the filesystem.
I don't know if you can recover this, but there might be a way someone can
suggest.
I have a question about the solution to solve the "Shard unassigned"
problem.
Here my setup, pretty simple:
1 NODE - 2 SHARDS / index - NO REPLICA
The situation is the following:
I manually moved all the content of /var/data/elasticsearch to
another location for backup need;
So I am able to start with a fresh cluster.
I restarted the service and everything was going fine until I realised
that all my Kibana dashboards were not available on the newly live
cluster.
I understood it was because of the previous move of the files, so I
simply moved the* kibana-int* index directory from the backup dir back to /var/data/elasticsearch
If it was so simple...
The new cluster showed the old kibana-int index as unassigned despite
restart of the service.
Questions:
How the new cluster can unassigned this index (does it see that it does
not created this index itself based on timestamp or other, and therefore
decide to not take it...) ?
How can I assign this index ? The only solution I know is to start
another node (so a 2 nodes cluster) hoping that the second node will take
the* kibana-int *and then move this index onto the first node.
Thanks a lot for sharing your experience on this matters, I am sure it
will help other.
Yeah, I agree.
It is kind of frustrating, as I know it is on the right place but ES gives
me still YELLOW !
I wish It was only a matter of modifying a single file in the index
directory.
It's probably because the cluster state doesn't know about the index even
though it can see it on the filesystem.
I don't know if you can recover this, but there might be a way someone can
suggest.
I have a question about the solution to solve the "Shard unassigned"
problem.
Here my setup, pretty simple:
1 NODE - 2 SHARDS / index - NO REPLICA
The situation is the following:
I manually moved all the content of /var/data/elasticsearch to
another location for backup need;
So I am able to start with a fresh cluster.
I restarted the service and everything was going fine until I realised
that all my Kibana dashboards were not available on the newly live
cluster.
I understood it was because of the previous move of the files, so I
simply moved the* kibana-int* index directory from the backup dir back
to /var/data/elasticsearch
If it was so simple...
The new cluster showed the old kibana-int index as unassigned despite
restart of the service.
Questions:
How the new cluster can unassigned this index (does it see that it
does not created this index itself based on timestamp or other, and
therefore decide to not take it...) ?
How can I assign this index ? The only solution I know is to start
another node (so a 2 nodes cluster) hoping that the second node will take
the* kibana-int *and then move this index onto the first node.
Thanks a lot for sharing your experience on this matters, I am sure it
will help other.
Best Regards,
Alex
--
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.
Does anyone has a definitive answer to this problem ?
Alex.
Le lundi 11 août 2014 19:11:02 UTC+2, AlexK a écrit :
Yeah, I agree.
It is kind of frustrating, as I know it is on the right place but ES
gives me still YELLOW !
I wish It was only a matter of modifying a single file in the index
directory.
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.