Re: Ideal hardware for NiFi

2018-09-14 Thread Mark Payne
e Witt > Sent: Friday, 14 September 2018 1:37 PM > To: dev@nifi.apache.org > Subject: Re: Ideal hardware for NiFi > > The ff disk needs to be the quickest disk and should have no other > contention just like a db trans log would request. > > The prov repo should also have

RE: Ideal hardware for NiFi

2018-09-13 Thread Phil H
Witt Sent: Friday, 14 September 2018 1:37 PM To: dev@nifi.apache.org Subject: Re: Ideal hardware for NiFi The ff disk needs to be the quickest disk and should have no other contention just like a db trans log would request. The prov repo should also have its pwn disk. The content repo can have

Re: Ideal hardware for NiFi

2018-09-13 Thread Joe Witt
The ff disk needs to be the quickest disk and should have no other contention just like a db trans log would request. The prov repo should also have its pwn disk. The content repo can have one or more physical disks. The best case is each repo is on physically separate disks/underlying storage.

Re: Ideal hardware for NiFi

2018-09-13 Thread Phil H
Hi joe, I moved the content and providence repositories off to two new disks, but it seems like the vast majority of the writes are still occurring on the disk where the flowfile and database repositories are. I note they don't appear to be able to be split across disks in the same way? On Fri,

Re: Ideal hardware for NiFi

2018-09-13 Thread Joe Witt
if they are physically seperate the diff should be quite noticable. On Thu, Sep 13, 2018, 7:36 PM Phil H wrote: > Potentially. We're looking to see how the multiple disks help before > committing to spending money on new hardware :) > > On Fri, 14 Sep 2018 at 10:48, Joe Witt wrote: > > > phil,

Re: Ideal hardware for NiFi

2018-09-13 Thread Phil H
Potentially. We're looking to see how the multiple disks help before committing to spending money on new hardware :) On Fri, 14 Sep 2018 at 10:48, Joe Witt wrote: > phil, > > as you add dirs it will just start using them. if you want to no longer > use the current dir it might be more

Re: Ideal hardware for NiFi

2018-09-13 Thread Joe Witt
phil, as you add dirs it will just start using them. if you want to no longer use the current dir it might be more involved. does that help? thanks On Thu, Sep 13, 2018, 4:36 PM Phil H wrote: > Follow up question - how do I transition to this new structure? Should I > shut down NiFi and

Re: Ideal hardware for NiFi

2018-09-13 Thread Phil H
Follow up question - how do I transition to this new structure? Should I shut down NiFi and move the contents of the legacy single directories into one of the new ones? For example: mv /usr/nifi/content_repository /nifi/repos/content-1 TIA Phil On Wed, 12 Sep 2018 at 06:15, Mark Payne wrote:

Re: Ideal hardware for NiFi

2018-09-11 Thread Phil H
Fantastic, thanks Mark On Wed, 12 Sep 2018 at 06:15, Mark Payne wrote: > Phil, > > For the content repository, you can configure the directory by changing > the value of > the "nifi.content.repository.directory.default" property in > nifi.properties. The suffix here, > "default" is the name of

Re: Ideal hardware for NiFi

2018-09-11 Thread Mark Payne
Phil, For the content repository, you can configure the directory by changing the value of the "nifi.content.repository.directory.default" property in nifi.properties. The suffix here, "default" is the name of this "container". You can have multiple containers by adding extra properties. So,

Re: Ideal hardware for NiFi

2018-09-11 Thread Phil H
Thanks Mark, this is great advice. Disk access is certainly an issue with the current set up. I will certainly shoot for NVMe disks in the build. How does NiFi get configured to span it's repositories across multiple physical disks? Thanks, Phil On Wed, 12 Sep 2018 at 01:32, Mark Payne wrote:

Re: Ideal hardware for NiFi

2018-09-11 Thread Mark Payne
Phil, As Sivaprasanna mentioned, your bottleneck will certainly depend on your flow. There's nothing inherent about NiFi or the JVM, AFAIK that would limit you. I've seen NiFi run on VM's containing 4-8 cores, and I've seen it run on bare metal on servers containing 96+ cores. Most often, I see

Re: Ideal hardware for NiFi

2018-09-10 Thread Phil H
Thanks for that. Sorry I should have been more specific - we have a flow running already on non-dedicated hardware. Looking to identify any limitations in NiFi/JVM that would limit how much parallelism it can take advantage of On Mon, 10 Sep 2018 at 14:32, Sivaprasanna wrote: > Phil, > > The