HA about janusGraph


Vayne Yang <vayne...@...>
 

Hi JanusGraph users,

    If I have a cassandra cluster with 4 machines, what should I configure the  'storage.hostname' as to make the most HA?

    This is my configuration now:  storage.hostname = 172.1.1.1, 172.1.1.2, 172.1.1.3, 172.1.1.4, but once A cassandra node is DOWN, the query to gremlin-server may be failed.

    Thank you for any hint and response!


HadoopMarc <bi...@...>
 

Hi Vayne Yang,
I am not sure this is the problem, but the spaces in your storage.hostname string look suspect.

HTH,    Marc

Op maandag 10 augustus 2020 om 04:57:33 UTC+2 schreef Vayne Yang:

Hi JanusGraph users,

    If I have a cassandra cluster with 4 machines, what should I configure the  'storage.hostname' as to make the most HA?

    This is my configuration now:  storage.hostname = 172.1.1.1, 172.1.1.2, 172.1.1.3, 172.1.1.4, but once A cassandra node is DOWN, the query to gremlin-server may be failed.

    Thank you for any hint and response!


Vayne Yang <vayne...@...>
 

hi Marc!
  Thanks for your reply!
  I am pretty sure the spaces are not the problems. e.g. https://docs.janusgraph.org/basics/example-config/#cassandra-remote configured in the same way. And after open the configure file with JanusGraphFactory, gremlin console also print 'standardjanusgraph[cql:[172.1.1.1, 172.1.1.2, 172.1.1.3, 172.1.1.4]]'.
  


On Monday, August 10, 2020 at 1:34:01 PM UTC+8, HadoopMarc wrote:
Hi Vayne Yang,
I am not sure this is the problem, but the spaces in your storage.hostname string look suspect.

HTH,    Marc

Op maandag 10 augustus 2020 om 04:57:33 UTC+2 schreef Vayne Yang:
Hi JanusGraph users,

    If I have a cassandra cluster with 4 machines, what should I configure the  'storage.hostname' as to make the most HA?

    This is my configuration now:  storage.hostname = 172.1.1.1, 172.1.1.2, 172.1.1.3, 172.1.1.4, but once A cassandra node is DOWN, the query to gremlin-server may be failed.

    Thank you for any hint and response!


HadoopMarc <bi...@...>
 

Hi Vayne Yang,

Can you eleborate on "the query may be failed"? Is it just a warning followed by an automatic new attempt or does the query produce no result? Can you add the stacktrace?

Best wishes,    Marc

Op maandag 10 augustus 2020 om 08:13:41 UTC+2 schreef Vayne Yang:

hi Marc!
  Thanks for your reply!
  I am pretty sure the spaces are not the problems. e.g. https://docs.janusgraph.org/basics/example-config/#cassandra-remote configured in the same way. And after open the configure file with JanusGraphFactory, gremlin console also print 'standardjanusgraph[cql:[172.1.1.1, 172.1.1.2, 172.1.1.3, 172.1.1.4]]'.
  

On Monday, August 10, 2020 at 1:34:01 PM UTC+8, HadoopMarc wrote:
Hi Vayne Yang,
I am not sure this is the problem, but the spaces in your storage.hostname string look suspect.

HTH,    Marc

Op maandag 10 augustus 2020 om 04:57:33 UTC+2 schreef Vayne Yang:
Hi JanusGraph users,

    If I have a cassandra cluster with 4 machines, what should I configure the  'storage.hostname' as to make the most HA?

    This is my configuration now:  storage.hostname = 172.1.1.1, 172.1.1.2, 172.1.1.3, 172.1.1.4, but once A cassandra node is DOWN, the query to gremlin-server may be failed.

    Thank you for any hint and response!


Vayne Yang <vayne...@...>
 

Hi Marc,
  Is that means, JanusGraph itself is HA right? If so, maybe the problem is that read-consistency-level=QUORUM  is too strict.
  Sorry I can't repeat the error immediately. if I run into this problem again, I will report here.
  Thanks for your help !


On Monday, August 10, 2020 at 2:43:00 PM UTC+8, HadoopMarc wrote:
Hi Vayne Yang,

Can you eleborate on "the query may be failed"? Is it just a warning followed by an automatic new attempt or does the query produce no result? Can you add the stacktrace?

Best wishes,    Marc

Op maandag 10 augustus 2020 om 08:13:41 UTC+2 schreef Vayne Yang:
hi Marc!
  Thanks for your reply!
  I am pretty sure the spaces are not the problems. e.g. https://docs.janusgraph.org/basics/example-config/#cassandra-remote configured in the same way. And after open the configure file with JanusGraphFactory, gremlin console also print 'standardjanusgraph[cql:[172.1.1.1, 172.1.1.2, 172.1.1.3, 172.1.1.4]]'.
  

On Monday, August 10, 2020 at 1:34:01 PM UTC+8, HadoopMarc wrote:
Hi Vayne Yang,
I am not sure this is the problem, but the spaces in your storage.hostname string look suspect.

HTH,    Marc

Op maandag 10 augustus 2020 om 04:57:33 UTC+2 schreef Vayne Yang:
Hi JanusGraph users,

    If I have a cassandra cluster with 4 machines, what should I configure the  'storage.hostname' as to make the most HA?

    This is my configuration now:  storage.hostname = 172.1.1.1, 172.1.1.2, 172.1.1.3, 172.1.1.4, but once A cassandra node is DOWN, the query to gremlin-server may be failed.

    Thank you for any hint and response!


Vayne Yang <vayne...@...>
 

Hi Marc,

  One more question please! When a query is sent to gremlin-server, what will the gremlin-server do with the storage backend? Is there any possible that the query is sent to a DOWN cassandra node which cause the query failed?

  Best wishes!


On Monday, August 10, 2020 at 2:43:00 PM UTC+8, HadoopMarc wrote:
Hi Vayne Yang,

Can you eleborate on "the query may be failed"? Is it just a warning followed by an automatic new attempt or does the query produce no result? Can you add the stacktrace?

Best wishes,    Marc

Op maandag 10 augustus 2020 om 08:13:41 UTC+2 schreef Vayne Yang:
hi Marc!
  Thanks for your reply!
  I am pretty sure the spaces are not the problems. e.g. https://docs.janusgraph.org/basics/example-config/#cassandra-remote configured in the same way. And after open the configure file with JanusGraphFactory, gremlin console also print 'standardjanusgraph[cql:[172.1.1.1, 172.1.1.2, 172.1.1.3, 172.1.1.4]]'.
  

On Monday, August 10, 2020 at 1:34:01 PM UTC+8, HadoopMarc wrote:
Hi Vayne Yang,
I am not sure this is the problem, but the spaces in your storage.hostname string look suspect.

HTH,    Marc

Op maandag 10 augustus 2020 om 04:57:33 UTC+2 schreef Vayne Yang:
Hi JanusGraph users,

    If I have a cassandra cluster with 4 machines, what should I configure the  'storage.hostname' as to make the most HA?

    This is my configuration now:  storage.hostname = 172.1.1.1, 172.1.1.2, 172.1.1.3, 172.1.1.4, but once A cassandra node is DOWN, the query to gremlin-server may be failed.

    Thank you for any hint and response!


HadoopMarc <bi...@...>
 

Hi Vayne Yang,

Query failing to a DOWN node because of a wrong node selection seems unlikely because JanusGraph uses the regular cassandra/cql drivers.

The similar thread below suggests to look into the gremlin-server logs for QUORUM warnings and to also configure the replication-factor on the janusgraph side.

HTH,    Marc

Op maandag 10 augustus 2020 om 10:44:00 UTC+2 schreef Vayne Yang:

Hi Marc,

  One more question please! When a query is sent to gremlin-server, what will the gremlin-server do with the storage backend? Is there any possible that the query is sent to a DOWN cassandra node which cause the query failed?

  Best wishes!


On Monday, August 10, 2020 at 2:43:00 PM UTC+8, HadoopMarc wrote:
Hi Vayne Yang,

Can you eleborate on "the query may be failed"? Is it just a warning followed by an automatic new attempt or does the query produce no result? Can you add the stacktrace?

Best wishes,    Marc

Op maandag 10 augustus 2020 om 08:13:41 UTC+2 schreef Vayne Yang:
hi Marc!
  Thanks for your reply!
  I am pretty sure the spaces are not the problems. e.g. https://docs.janusgraph.org/basics/example-config/#cassandra-remote configured in the same way. And after open the configure file with JanusGraphFactory, gremlin console also print 'standardjanusgraph[cql:[172.1.1.1, 172.1.1.2, 172.1.1.3, 172.1.1.4]]'.
  

On Monday, August 10, 2020 at 1:34:01 PM UTC+8, HadoopMarc wrote:
Hi Vayne Yang,
I am not sure this is the problem, but the spaces in your storage.hostname string look suspect.

HTH,    Marc

Op maandag 10 augustus 2020 om 04:57:33 UTC+2 schreef Vayne Yang:
Hi JanusGraph users,

    If I have a cassandra cluster with 4 machines, what should I configure the  'storage.hostname' as to make the most HA?

    This is my configuration now:  storage.hostname = 172.1.1.1, 172.1.1.2, 172.1.1.3, 172.1.1.4, but once A cassandra node is DOWN, the query to gremlin-server may be failed.

    Thank you for any hint and response!


Vayne Yang <vayne...@...>
 

Hi Marc,
  
  I see, Thank you so much!

On Monday, August 10, 2020 at 5:21:38 PM UTC+8, HadoopMarc wrote:
Hi Vayne Yang,

Query failing to a DOWN node because of a wrong node selection seems unlikely because JanusGraph uses the regular cassandra/cql drivers.

The similar thread below suggests to look into the gremlin-server logs for QUORUM warnings and to also configure the replication-factor on the janusgraph side.

HTH,    Marc

Op maandag 10 augustus 2020 om 10:44:00 UTC+2 schreef Vayne Yang:
Hi Marc,

  One more question please! When a query is sent to gremlin-server, what will the gremlin-server do with the storage backend? Is there any possible that the query is sent to a DOWN cassandra node which cause the query failed?

  Best wishes!


On Monday, August 10, 2020 at 2:43:00 PM UTC+8, HadoopMarc wrote:
Hi Vayne Yang,

Can you eleborate on "the query may be failed"? Is it just a warning followed by an automatic new attempt or does the query produce no result? Can you add the stacktrace?

Best wishes,    Marc

Op maandag 10 augustus 2020 om 08:13:41 UTC+2 schreef Vayne Yang:
hi Marc!
  Thanks for your reply!
  I am pretty sure the spaces are not the problems. e.g. https://docs.janusgraph.org/basics/example-config/#cassandra-remote configured in the same way. And after open the configure file with JanusGraphFactory, gremlin console also print 'standardjanusgraph[cql:[172.1.1.1, 172.1.1.2, 172.1.1.3, 172.1.1.4]]'.
  

On Monday, August 10, 2020 at 1:34:01 PM UTC+8, HadoopMarc wrote:
Hi Vayne Yang,
I am not sure this is the problem, but the spaces in your storage.hostname string look suspect.

HTH,    Marc

Op maandag 10 augustus 2020 om 04:57:33 UTC+2 schreef Vayne Yang:
Hi JanusGraph users,

    If I have a cassandra cluster with 4 machines, what should I configure the  'storage.hostname' as to make the most HA?

    This is my configuration now:  storage.hostname = 172.1.1.1, 172.1.1.2, 172.1.1.3, 172.1.1.4, but once A cassandra node is DOWN, the query to gremlin-server may be failed.

    Thank you for any hint and response!