Ashish, The system tries to spread the primary and secondary copies on all the available nodes. I guess you have 12 nodes with the partitioned region created before the export; and every node will have few primary buckets and few secondary buckets created after the export.
>> why it is creating copies on all data nodes? I am not clear what you meant with this; can you elaborate on this. -Anil. On Tue, Jul 23, 2019 at 1:23 PM aashish choudhary < [email protected]> wrote: > Not sure if parallel option is available in geode version 1.2 for > partitioned regions so it means it will export data serially to a single > file? > > Are you saying that Partitioned Region and its copies are created on all > 12 nodes (in the cluster). Yes > > Are you expecting it all the buckets to be created on one member? No I was > expecting it to load balanced across nodes as you said but if I specify > total 3 copies why it is creating copies on all data nodes? > > > > With best regards, > Ashish > > On Wed, Jul 24, 2019, 1:24 AM Anilkumar Gingade <[email protected]> > wrote: > >> For partitioned region, the data can exported and imported in two ways, >> serial and parallel. >> You can find more info on this at: >> >> https://geode.apache.org/docs/guide/16/tools_modules/gfsh/command-pages/export.html#topic_263B70069BFC4A7185F86B3272011734 >> >> https://geode.apache.org/docs/guide/16/tools_modules/gfsh/command-pages/import.html#topic_jw2_2ld_2l >> >> Read for "export data" and "Import data" section. >> >> For a large region; the parallel export will help; as it tries to export >> the data in parallel. You can find more on performance impact at: >> >> https://geode.apache.org/docs/guide/11/managing/cache_snapshots/using_cache_and_region_snapshots.html >> >> >> But when we imported the data to that partitioned region copy is >> created on all cache nodes. No idea why? >> Are you saying that Partitioned Region and its copies are created on all >> 12 nodes (in the cluster). Are you expecting it all the buckets to be >> created on one member? >> >> When you import data, the data gets spread across all the nodes which >> hosts that partitioned region; and the amount of data is balanced across >> all nodes. You will see Primary buckets getting created across the nodes >> and secondary buckets on other nodes. >> >> -Anil. >> >> >> >> >> >> >> >> >> >> On Tue, Jul 23, 2019 at 9:37 AM aashish choudhary < >> [email protected]> wrote: >> >>> Thanks Anil. >>> >>> Need few more confirmation >>> >>> >>> >>> We exported/imported the data for replicated region and it worked fine. >>> For partitioned region does the export command export all data from other >>> nodes to single file? >>> >>> >>> >>> For partitioned region before importing we created region with redundant >>> copies of 2 means we have total 3 copies including 2 secondary and one >>> primary. But when we imported the data to that partitioned region copy is >>> created on all cache nodes. No idea why? But describe region shows correct >>> region count as source cluster. Its just that copy seems to be created on >>> all 12 nodes >>> >>> >>> >>> Can someone confirm what total bucket size means in show metrics command? >>> >>> >>> >>> What performance impact would be there on source cluster if we do export >>> of big partitioned region having millions of records? >>> >>> With best regards, >>> Ashish >>> >>> On Tue, Jul 23, 2019, 6:14 AM Anilkumar Gingade <[email protected]> >>> wrote: >>> >>>> Yes, import and export. >>>> >>>> Alter region does not support changing the redundancy level. You can >>>> create region with different redundant copies and import data into it; it >>>> will work. >>>> >>>> -Anil. >>>> >>>> >>>> On Mon, Jul 22, 2019 at 5:33 PM aashish choudhary < >>>> [email protected]> wrote: >>>> >>>>> You mean to say import/export? >>>>> >>>>> With that enabling redundancy will still work using alter region after >>>>> importing data.? >>>>> >>>>> With best regards, >>>>> Ashish >>>>> >>>>> On Tue, Jul 23, 2019, 2:38 AM Anilkumar Gingade <[email protected]> >>>>> wrote: >>>>> >>>>>> Hi Ashish, >>>>>> >>>>>> If you are planning to change the cluster size and region >>>>>> configuration; backup and restore will not work. You need to take data >>>>>> snapshot and load the snapshot. >>>>>> >>>>>> -Anil. >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> On Mon, Jul 22, 2019 at 11:39 AM aashish choudhary < >>>>>> [email protected]> wrote: >>>>>> >>>>>>> Hi, >>>>>>> >>>>>>> Need some input on below scenario. We are planning to do a hardware >>>>>>> upgrade from virtual machine to physical hardware. We will be using same >>>>>>> geode version in physical machines. >>>>>>> >>>>>>> Questions >>>>>>> >>>>>>> To avoid data ingestion on new physical hardware can we take backup >>>>>>> from vms and restore them without any issue? >>>>>>> >>>>>>> In current VM setup we don't have redundancy enabled so can we >>>>>>> enable redundancy using alter region command after restoring the backup? >>>>>>> >>>>>>> In VM we are running with 3 data nodes and 3 locators. In physical >>>>>>> machines we will be running with running with 12 nodes and 3 locators( >>>>>>> total 6 machines one node will run 2 data nodes). Will this setup create >>>>>>> any problems while restoring the backup? >>>>>>> >>>>>>> >>>>>>> With best regards, >>>>>>> Ashish >>>>>>> >>>>>>
