JanusGraph•4w ago

Janusgraph bigtable rows exceeds the limit 256MiB when exported via Dataflow in Parquet format

Hi team, Currently, we are using Janusgraph with Bigtable as the storage backend. And we wanted to export the data out of Bigtable using Dataflow in a Parquet format to cloud storage. But during the process it failed because some of the rows size too large that exceeds the limit with the following error messages: See attachment We have talked with GCP support if there is a workaround for this and they suggest to change the GC policy of the columns in the table. But since the rows and columns structure are created and managed directly by Janusgraph, we have concern that if we modify/change the GC policy, it might corrupt the data. Our question is, is there a way to configure the size of the rows in janusgraph? Or is it possible to configure the GC policy directly from Janusgraph? Do let me know if I posted this in a wrong section. Column families that have large row size:

risk-serving-bt-feature-engine 

family {
  name: "l"
  locality_group: "user_flash"
  administrator: "chubby!mdb/cloud-bigtable-internal-bigtable-administrators-prod"
  administrator: "chubby!user/cloud-bigtable"
  writer: "chubby!mdb/cloud-bigtable-internal-bigtable-writers-prod"
  writer: "chubby!user/cloud-bigtable"
  reader: "chubby!mdb/cloud-bigtable-internal-bigtable-readers-prod"
  reader: "chubby!user/cloud-bigtable"
  gcexpr: "(age() > 604800000000 || version() > 1)"
}

risk-serving-bt-feature-engine 

family {
  name: "l"
  locality_group: "user_flash"
  administrator: "chubby!mdb/cloud-bigtable-internal-bigtable-administrators-prod"
  administrator: "chubby!user/cloud-bigtable"
  writer: "chubby!mdb/cloud-bigtable-internal-bigtable-writers-prod"
  writer: "chubby!user/cloud-bigtable"
  reader: "chubby!mdb/cloud-bigtable-internal-bigtable-readers-prod"
  reader: "chubby!user/cloud-bigtable"
  gcexpr: "(age() > 604800000000 || version() > 1)"
}

janusgraph version: 0.6.4 storage backend: bigtable

message.txt

0 Replies

No replies yetBe the first to reply to this messageJoin

Gaming

Programming

Janusgraph bigtable rows exceeds the limit 256MiB when exported via Dataflow in Parquet format