[
https://issues.apache.org/jira/browse/SOLR-10265?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15931336#comment-15931336
]
Erick Erickson commented on SOLR-10265:
---------------------------------------
[~noble.paul] Interesting. You're essentially replacing the overseer zknodes
with an in-memory queue, right?
I do wonder about sequencing. The order of operations could get switched,
although I'm not sure it matters if there's some sort of optimistic concurrency
provisions (or maybe even if not). Imagine request1 and request2 in that order,
the overseer dies. Depending on the timing of the retry logic, the clients
could well submit them in reverse order....
Another thought I had was how often we actually need to write the state out.
Imagine 100 updates for the same state.json file. Is there any good reason
those updates couldn't all be written at once? The algorithm would look
something like
get N messages from the queue
Apply all the changes
write out all affected nodes to ZK.
Hmmm, in that case, though, any watchers expecting an inermediate transition
state wouldn't see it. Imagine node goes straight from "down" to "active". That
is all the intermediate state changes are collapsed into a single write. Would
that matter? In the situation we're talking about here this would certainly cut
down a lot of chatter.
Musings on Saturday morning.
> Overseer can become the bottleneck in very large clusters
> ---------------------------------------------------------
>
> Key: SOLR-10265
> URL: https://issues.apache.org/jira/browse/SOLR-10265
> Project: Solr
> Issue Type: Improvement
> Security Level: Public(Default Security Level. Issues are Public)
> Reporter: Varun Thacker
>
> Let's say we have a large cluster. Some numbers:
> - To ingest the data at the volume we want to I need roughly a 600 shard
> collection.
> - Index into the collection for 1 hour and then create a new collection
> - For a 30 days retention window with these numbers we would end up wth
> ~400k cores in the cluster
> - Just a rolling restart of this cluster can take hours because the overseer
> queue gets backed up. If a few nodes looses connectivity to ZooKeeper then
> also we can end up with lots of messages in the Overseer queue
> With some tests here are the two high level problems we have identified:
> 1> How fast can the overseer process operations:
> The rate at which the overseer processes events is too slow at this scale.
> I ran {{OverseerTest#testPerformance}} which creates 10 collections ( 1 shard
> 1 replica ) and generates 20k state change events. The test took 119 seconds
> to run on my machine which means ~170 events a second. Let's say a server can
> process 5x of my machine so 1k events a second.
> Total events generated by a 400k replica cluster = 400k * 4 ( state changes
> till replica become active ) = 1.6M / 1k events a second will be 1600 minutes.
> Second observation was that the rate at which the overseer can process events
> slows down when the number of items in the queue gets larger
> I ran the same {{OverseerTest#testPerformance}} but changed the number of
> events generated to 2000 instead. The test took only 5 seconds to run. So it
> was a lot faster than the test run which generated 20k events
> 2> State changes overwhelming ZK:
> For every state change Solr is writing out a big state.json to zookeeper.
> This can lead to the zookeeper transaction logs going out of control even
> with auto purging etc set .
> I haven't debugged why the transaction logs ran into terabytes without taking
> into snapshots but this was my assumption based on the other problems we
> observed
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org
