Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
    Filippo De Luca
    @filosganga
    The UserProfileProcessor is just started one for each node, and keep in sync with PubSubMediator
    Martin Jöhren
    @matlockx
    Yepp
    Martin Jöhren
    @matlockx
    Same persistenceId will override events for sure. I wouldn't do that. Regarding Cassandra version: I migrated our cluster last week from 2.1 to 2.2 without any problems and later on I migrated 10 services from akka 2.3 to 2.4 also without any problems. I used a small python script for it. It's just time consuming but we went through it without downtime. Our services are running at least in pairs and they are completely independent, so different persistence ids. That is why we could do it without any downtime and risk.
    Patrik Nordwall
    @patriknw
    +1 "it won't end well", the persistenceId must be unique, single writer principle
    Mark
    @Fristi
    Is there any good GUI cassandra tool to take a peek inside a cassandra db on osx?
    Martin Jöhren
    @matlockx
    I don't know any but I would suggest to use cqlsh instead. If you are familiar with sql it's not quite far away and you can easily explore the database. Provided help is good and it has auto completion
    Mark
    @Fristi
    Yea came across that! SQL is no problem, but a GUI db tool can be handy for basic ops ;)
    Mark
    @Fristi
    Hmm does cqlsh already support cassandra 3x? I am getting this;
    $ cqlsh --version
    cqlsh 5.0.1
    $ cqlsh
    Connection error: ('Unable to connect to any servers', 
    {'127.0.0.1': InvalidRequest(u'code=2200 [Invalid query] message="unconfigured table schema_keyspaces"',)})
    Martin Jöhren
    @matlockx
    Yepp. Seems that you have no local Cassandra running which is listening on localhost port 9042
    Regarding incompatibility: as far as I know you can only run in python dependency problems, e.g. Too old python version
    Simon
    @simon-lam

    Hello all! I'm a bit of a newer Akka user and have been very interested in understanding how Persistence Queries work and have been using akka-persistence-cassandra locally as a learning exercise.

    One thing I still have not been able to fully understand after looking through the code is the approach for making sure NEW journal events are propagated from the write side to the read journal (per the docs, a query plugin implementing functions like allPersistenceIds should essentially be a stream that does not close and emits new events).

    Would someone be able to point be in the right direction?

    Patrik Nordwall
    @patriknw
    @simon-lam It queries Cassandra periodically to look for new events. There is no push mechanism from Cassandra to a client (yet) so the only thing we can do is to poll.
    Simon
    @simon-lam
    ahhh i see now
    thanks
    Martin Jöhren
    @matlockx
    Hi guys, we had today a real bad issue after the migration to the new schema. We migrated the data before Christmas for several services. All of them had events in the millions. Unfortunately we missed that the default partition size shrinked from 5000000 to 500000.. Just a zero.... Problem now was that at first we had no issues. Events got persisted, apps were running fine. But after a restart we 'lost' the new events that have been created after the migration. Problem was, that they got a new partition number with skipping several numbers. So the recovery process found events for partition 0 but the next partition number was missing and therefore it stopped. I would suggest to point that out more clearly in the readme. I like to provide that as pr, question is where (akka or krasserm) since typesafe took over maintenance.
    Patrik Nordwall
    @patriknw
    I'm sorry to hear that. We will absolutely be more strict about changes and documenting migrations, especially after 1.0 is released. Help is appreciated. The akka repository is not active yet, so please PR krasserm.
    Martin Jöhren
    @matlockx
    Appreciated. Will do the pr. Just to be a bit more precise: we have not lost the events, they just haven't been considered during recovery. We fixed that by manually increasing the partition number for a couple of events so that the recovery process was able to read in all events. Business impact on our side was, well, not that good but also not that bad. Hope no one else gets into that.
    Patrik Nordwall
    @patriknw
    Thanks for sharing
    Christopher Batey
    @chbatey
    for future uprades your app would fail to start as that config option is persisted and checked that it has not changed
    countfloyd
    @countfloyd
    @patriknw as far as you know are there plans for a push mechanism from Cassandra to deliver new events to the query side? That would be the bees knees.
    Patrik Nordwall
    @patriknw
    @countfloyd I have heard about that, but I don't know much about it. That will indeed be very interesting. Do you have any pointers?
    Patrik Nordwall
    @patriknw
    thanks
    countfloyd
    @countfloyd
    @patriknw I think if you look at the leveldb query driver that comes with akka, they implement an immediate push for new messages without going to the database.
    Patrik Nordwall
    @patriknw
    Yeah, I did that :-)
    it's easier there, because things are local. Push from Cassandra would be great.
    Martin Grotzke
    @magro
    Hi, I've one question: is the new persistence queries implementation only available for C* 3.x? According to the docs this seems to be the case.
    We need backpressure with event replay for a single persistenceId (this would probably be eventsByPersistenceId) - is this somehow possible with the old event replay mechanism? AFAICS this is not directly possible (without "hacking" the CassandraJournal)...
    Using C* 3.x in production is a thing of the future, as soon as there are some minor releases out - this is what Datastax recommends... Therefore we cannot use 3.x ATM...
    Patrik Nordwall
    @patriknw
    @magro the queries are only available for the 3.x version
    we don't have the bandwidth to backport everything, and the eventsByTag query is using new features in C* 3.x
    hopefully C* 3.x will become enough production ready soon
    Martin Grotzke
    @magro
    Thanks, @patriknw!
    Do you think it's much effort to backport simple queries?
    Martin Grotzke
    @magro
    Btw, are materialized views now used for e.g.
    events by Tag?
    Assuming this is the case, are they also used /created/active even if I don't want to use such complex queries?
    Basically I'd like to avoid the costs (additional load) that come with materialized views.
    Patrik Nordwall
    @patriknw
    @magro you can try to backport, I see no reason why it would not work (apart from eventsByTag). If you do the backport I don't mind cutting a release. Note that the new home is https://github.com/akka/akka-persistence-cassandra and there the target branch is cassandra-2.x
    @magro yes, materialized view is used for eventsByTag. It should not add much overhead if events are not tagged, but I can add a disable config flag. akka/akka-persistence-cassandra#22
    Martin Grotzke
    @magro
    @patriknw I like the disabled flag, just added a comment...
    Let's see how our new story will be prioritized, then backporting would probably make sense for us.
    Another thing: On https://github.com/akka/akka-persistence-cassandra I checked the gitter channel (#akka) and decided to ask here :-) Maybe it makes sense to use dedicated channels for some akka extensions like apc?
    Patrik Nordwall
    @patriknw
    @/all I have created a new gitter channel in the akka repo: https://gitter.im/akka/akka-persistence-cassandra Please join!
    Martin Grotzke
    @magro
    Yeah! :-)
    jaksky
    @jaksky
    Hello, I have issues during the application startup that first message after startup its events doesn't get persisted into journal. Subsequent one are ok. Any useful logging can help me to nail it down?
    anybody @here can help me?
    Martin Krasser
    @krasserm
    @jaksky This repo is no longer maintained. Please use https://github.com/akka/akka-persistence-cassandra instead. Thanks!
    Srinivas
    @Sriniva63328880_twitter
    Hi all
    sumesh92
    @sumesh92
    Hello Folks, Is it best practice to update and delete in Cassandra database table? I am a newbie on Cassandra. Please help me