Re: Is re-election/assignment of the master node possible?
The load issue affecting master detection / election shouldn't happen if you have dedicated masters... At least it is with 0.90.x ( with my limited knowledge of ES implementation details, there seems to be a lock or priority issue when serving large # of requests (http / thrift) , affecting cluster / metadata updates... I would think these metadata tasks ought to take priority in some cases over queries... ) On 26/11/2014 6:11 pm, Erik theRed j.e.redd...@gmail.com wrote: Thanks, Nik - There's no data on the node so it sounds like master reelection should fail over fairly quickly. On Wednesday, November 26, 2014 2:58:43 PM UTC-6, Nikolas Everett wrote: On Wed, Nov 26, 2014 at 3:47 PM, Erik theRed j.e.r...@gmail.com wrote: Is there any notion triggering a re-election of the master node? I'm currently running 1.2.4, and I have an instance that is scheduled for retirement (my favorite!) and it just so happens that it's my master node. What can I do to avoid the dreaded RED state? Is there some mechanism that can allow me to re-assign the current master to one of the other available two dedicated master nodes so I can reboot the current master? Move all the shards off of the node using allocation include/exclude settings. If you shoot the master one of the other master eligible nodes will take over quickly and there won't be any interruptions. I ask because I'm a bit gun-shy due to my experience when an elected master node has gone unresponsive (before I created dedicated masters) due to excessive HTTP connections, master re-election seemed to never occur and everything comes crumbling down. I've never had that problem. My cluster is pretty small though - only 31 nodes. Nik -- 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+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/elasticsearch/b9506885-e321-4abe-b1c2-db0d802b07ec%40googlegroups.com https://groups.google.com/d/msgid/elasticsearch/b9506885-e321-4abe-b1c2-db0d802b07ec%40googlegroups.com?utm_medium=emailutm_source=footer . For more options, visit https://groups.google.com/d/optout. -- 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+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/elasticsearch/CACj2-4J8Ycj07jhHdX71JJjAHW0_ZALMH9mPUcW1__sF_1NagA%40mail.gmail.com. For more options, visit https://groups.google.com/d/optout.
Is re-election/assignment of the master node possible?
Is there any notion triggering a re-election of the master node? I'm currently running 1.2.4, and I have an instance that is scheduled for retirement (my favorite!) and it just so happens that it's my master node. What can I do to avoid the dreaded RED state? Is there some mechanism that can allow me to re-assign the current master to one of the other available two dedicated master nodes so I can reboot the current master? I ask because I'm a bit gun-shy due to my experience when an elected master node has gone unresponsive (before I created dedicated masters) due to excessive HTTP connections, master re-election seemed to never occur and everything comes crumbling down. Thanks! - Erik -- 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+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/elasticsearch/80d7ec69-40e6-46d6-8de1-159d6117a7c7%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
Re: Is re-election/assignment of the master node possible?
On Wed, Nov 26, 2014 at 3:47 PM, Erik theRed j.e.redd...@gmail.com wrote: Is there any notion triggering a re-election of the master node? I'm currently running 1.2.4, and I have an instance that is scheduled for retirement (my favorite!) and it just so happens that it's my master node. What can I do to avoid the dreaded RED state? Is there some mechanism that can allow me to re-assign the current master to one of the other available two dedicated master nodes so I can reboot the current master? Move all the shards off of the node using allocation include/exclude settings. If you shoot the master one of the other master eligible nodes will take over quickly and there won't be any interruptions. I ask because I'm a bit gun-shy due to my experience when an elected master node has gone unresponsive (before I created dedicated masters) due to excessive HTTP connections, master re-election seemed to never occur and everything comes crumbling down. I've never had that problem. My cluster is pretty small though - only 31 nodes. Nik -- 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+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/elasticsearch/CAPmjWd2qctbbzPgH4iKTw1TvRymCL125zXymAeFUo9%3D84COLXg%40mail.gmail.com. For more options, visit https://groups.google.com/d/optout.
Re: Is re-election/assignment of the master node possible?
Thanks, Nik - There's no data on the node so it sounds like master reelection should fail over fairly quickly. On Wednesday, November 26, 2014 2:58:43 PM UTC-6, Nikolas Everett wrote: On Wed, Nov 26, 2014 at 3:47 PM, Erik theRed j.e.r...@gmail.com javascript: wrote: Is there any notion triggering a re-election of the master node? I'm currently running 1.2.4, and I have an instance that is scheduled for retirement (my favorite!) and it just so happens that it's my master node. What can I do to avoid the dreaded RED state? Is there some mechanism that can allow me to re-assign the current master to one of the other available two dedicated master nodes so I can reboot the current master? Move all the shards off of the node using allocation include/exclude settings. If you shoot the master one of the other master eligible nodes will take over quickly and there won't be any interruptions. I ask because I'm a bit gun-shy due to my experience when an elected master node has gone unresponsive (before I created dedicated masters) due to excessive HTTP connections, master re-election seemed to never occur and everything comes crumbling down. I've never had that problem. My cluster is pretty small though - only 31 nodes. Nik -- 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+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/elasticsearch/b9506885-e321-4abe-b1c2-db0d802b07ec%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.