Disabling index stale transactions

35 views
Skip to first unread message

Collin Scangarella

unread,
Jul 21, 2016, 2:12:55 PM7/21/16
to Aurelius
I'm trying to disable an index in order to remove it. However, when I try to disable it I receive a stale transaction error on mgmt.commit():

gremlin> mgmt = graph.openManagement()
==>com.thinkaurelius.titan.graphdb.database.management.ManagementSystem@1f15e689
gremlin> index = mgmt.getGraphIndex('expectedTypeuuidAndQueryType')
==>expectedTypeuuidAndQueryType
gremlin> mgmt.updateIndex(index, SchemaAction.DISABLE_INDEX).get()
==>null
gremlin> mgmt.commit()
==>null
gremlin> 18:04:50.154 [Thread-10] ERROR c.t.t.g.d.m.ManagementLogger - Evicted [3@ac1f21d629900-ip-172-31-33-2141] from cache but waiting too long for transactions to close. Stale transaction alert on: [standardtitantx[0x188a5fc2], standardtitantx[0x75b6dd5b]]

Any subsequent calls to await the updated status fail:

gremlin> ManagementSystem.awaitGraphIndexStatus(graph, 'expectedTypeuuidAndQueryType').status(SchemaStatus.DISABLED).call()
==>GraphIndexStatusReport[success=false, indexName='expectedTypeuuidAndQueryType', targetStatus=DISABLED, notConverged={expectedType=INSTALLED, uuidAndQueryType=INSTALLED}, converged={}, elapsed=PT1M0.086S]

Does anyone know how I can resolve this?

Erik Mocný

unread,
Jul 22, 2016, 8:56:23 AM7/22/16
to Aurelius
Hi Colin,

try graph.tx.rollback() before mgmt = graph.openManagement()

Dňa štvrtok, 21. júla 2016 20:12:55 UTC+2 Collin Scangarella napísal(-a):
Reply all
Reply to author
Forward
0 new messages