i deploy elasticsearch cluster in azure. found articles explaining how configure vms , looks pretty easy. use vm "hard disk" store data.
i prefer separate vm operating system , elasticsearch installation data storage. upgrade vms bigger ones if needed or use more storage independtly. , how should configure it? little bit lost different storage options in azure. suppose need virtual "hard disk" elasticsearch. thank answer or thoughts.
i little confused realised needed. when spin new vm, 2 drives present on it.
the first drive relatively small os drive. persisted cloud storage virtual disk (vhd) stored in cloud storage container specified, or automatically created when started vm. changes (e.g., install elasticsearch) make disk saved. however, since drive relatively small, there isn't space available - it's not best practice store application data on it.
the second drive temporary local storage provided vm (e.g., 100gb ssd d2 vm). drive situated in physical machine itself, it's typically fast there no network latency access it. however, intended temp working data because transient. is, if vm gets migrated physical server in cloud - perhaps after scaling or hardware failure - lose data on because resides inside 1 physical machine only. not news if contained index!
the solution therefore create disk stored in cloud (just aforementioned os disk), index data stored separately physical machine hosting vm @ 1 moment in time. if vm moved physical machine, local storage wiped, additional disk containing index re-attached in previous state, ready es continue.
to add new/empty disk existing linux vm following tutorial: https://azure.microsoft.com/en-us/documentation/articles/virtual-machines-linux-how-to-attach-disk/. once disk added, need go elasticsearch.yml file , set path.data wherever choose mount additional disk.
Comments
Post a Comment