cannot update any node

[ismaeladra]

i have a neo4j casual cluster with 3 nodes i can't update nodes attributes or create new relation any query will last forever and locks will still active

**Neo4j Version:3.4.12
Operating System: centos 7.5

[jexp]

How do you try to update data?
Do you use a driver with the bolt+routing protocol?
And do you use a WRITE transaction/session?

[ismaeladra]

i tried three ways

1. neo4j browser
2. cypher shell
3. python py2neo with bolt+routing

[jexp]

What does your query look like and are there any messages in the debug.log files of the cluster member.

Are you sure the cluster has formed? Please check that first.

[ismaeladra]

my query : match (u:users) where u.user_id = 100 set u.username = 'test'
cluster is formed
and there are no error messages or anything suspicous in debug.log

[jexp]

And your db is empty to start with?

or do you already have users? If you do, make sure to have a constraint like:

create constraint on (u:users) assert u.user_id is unique

[jexp]

Can you share the output of call dbms.listQueries() and call dbms.listActiveLocks('your-query-id')

[ismaeladra]

my db is not empty and i already have users and i have this constraint created before

listqueries.xlsx

queryId	username	metaData	query	parameters	planner	runtime	indexes	startTime	protocol	clientAddress	requestUri	status	resourceInformation	activeLockCount	elapsedTimeMillis	cpuTimeMillis	waitTimeMillis	idleTimeMillis	allocatedBytes	pageHits	pageFaults
query-4389	SocialAdmin	{}	call dbms.listQueries()	{}	procedure	procedure	[]	2019-03-27T14:15:02.508Z	bolt	10.32.192.18:13309	10.0.31.61:7687	running	{}	0	4	null	0	null	null	0	0
query-3318	SocialAdmin	{}	match (u:wo_users) where u.user_id = '1050' set u.country_id = 209	{}	idp	slotted	[{propertyKey:user_id,identifier:u,indexType:SCHEMA INDEX,label:wo_users,labelId:1,entityType:NODE}]	2019-03-27T10:35:03.675Z	bolt	10.32.192.18:5510	10.0.31.61:7687	waiting	{waitTimeMillis:13198844,lockMode:EXCLUSIVE,resourceType:NODE,resourceIds:[3491]}	2	13198847	null	13198844	null	null	2	0

active locks

╒════════╤══════════════╤══════════════════╕
│ "mode" │"resourceType" │"resourceId" │
╞════════╪══════════════╪══════════════════╡
│ "SHARED"│"INDEX_ENTRY" │460640061450813690 │
├────────┼──────────────┼──────────────────┤
│ "SHARED"│"LABEL" │1 │
└────────┴──────────────┴──────────────────┘

[jexp]

Can you show the output of CALL dbms.cluster.overview()

[jexp]

Also if possible the status of your indexes:
call db.indexes

Do you have any other plugins / extensions installed/running that could affect this? E.g. some tx-event-handlers?

[jexp]

You could also try https://halin.graphapp.io and see if there is anything unusal.

[ismaeladra]

cluster overview result is :

result of call db.indexes :
indexes.xlsx

i don't have any plugin and i can't use this monotring tool becuase the db is not connected to internet and its inside my company intranet
can i user another monotirng tool like graphite ?

[jexp]

You can also install halin for offline use with Neo4j desktop.

[martinfurmanski]

@ismaeladra1 This sounds very odd and possibly hard to diagnose for a regular end user, but you could try to get a thread dump and see if there are "bolt workers" actively processing the queries that you are submitting, and if so how they are blocked.

When getting the thread dump you have to do that on the instance to which the query was routed. To figure that out either you can use direct bolt connections (no routing) or use the WRITE-mode and check on the leader node.

Perhaps using the Java driver directly and enabling the most detailed level of logging will yield some further insights.
https://neo4j.com/docs/api/java-driver/1.7/org/neo4j/driver/v1/Logging.html

Maybe that is possible with cypher-shell or browser as well easily, but it is a bit out of my area to know in detail.

Also some more general typical advice: use latest versions of server, of driver, try restarting, check your configurations, network connectivity, environment, etc.

[rmgpinto]

I have the same problem, did you solve this?