Hello,
using v2.1.13
I imported database from v2.1.0 and wanted to setup cluster with 3 nodes.
On each node restart( i tested to see how it works) it flushed whole
database and syncs it back from one node(even though there were no inserts
as it is not live yet) and it syncs 1MB of compressed data per second. for
48GB db(8GB compressed) so it takes ages every time. Is there a way to fix
that?
default-distribuded-db.config.json:
{
"autoDeploy": true,
"hotAlignment": false,
"executionMode": "undefined",
"readQuorum": 1,
"writeQuorum": 2,
"failureAvailableNodesLessQuorum": false,
"readYourWrites": true,
"servers": {
"*": "master"
},
"clusters": {
"internal": {
},
"index": {
},
"*": {
"servers": ["<NEW_NODE>"]
}
}
}
Servers are connected via tcp-ip in hazelcast. there is gigabit link
between servers and all of them are new machines with 64G ram and ssd
storage.
Also we tried to contain orientdb in 32G ram and on each reindex it takes
pretty much all the ram available. we tried lots of different
heap/diskcache combos but no luck. any recommendations on this?
Also as this setup is pretty much awful and it started as distributed setup
then was moved to single node there is a lot of clusters printed in "list
clusters" should i freeze/remove ones that are not used?
i tried to repair database but no luck:
orientdb {db=Events}> repair database
Repair of graph 'remote:localhost/Events' is started ...
Scanning 6596654 edges...
Error: java.lang.NullPointerException
any suggestions here are more then welcome. things liek best setup
scenario, best number of nodes. i thought about sharding maybe to
distribute indexes
Thanks!
--
---
You received this message because you are subscribed to the Google Groups
"OrientDB" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
For more options, visit https://groups.google.com/d/optout.
