actordb/actordb

lixiaolong1997

@lixiaolong1997

@SergejJurecko ,i issued a problem,can you check it ? please

Tristan Sloughter

@tsloughter

checking out actordb again since it was on HN :), noticed teh issues on github still say there is a new iteration that shares no code with the existing actordb in the works a year ago. is that still the case and still not public?

James Rasmussen

@heyitsjames

^ Same question as Tristan. We'd love to hear an update of your guys' progress. This seems like a potential solution to a few problems at my current company, and I'd love to hear how things are coming along.

Tristan Sloughter

@tsloughter

I'm also just interested in if you've moved to rebar3/relx and happy to help with that :)

Sergej Jurečko

@SergejJurecko

Erlang version will stay the main one. The other codebase is more of an experiment for now.

Tristan Sloughter

@tsloughter

oh, is it not Erlang?

Sergej Jurečko

@SergejJurecko

Mixture of rust and c

Tristan Sloughter

@tsloughter

Tristan Sloughter

@tsloughter

I hope to look at it in the context of working on https://github.com/spacetime-iot/erleans and this from Orleans https://www.microsoft.com/en-us/research/uploads/prod/2018/04/ICDE2018-Keynote.pdf

Fear Gan Ainm

@Verace

Hi - I'm really interested in your project. I've also been looking at CockroachDB and TiDB. I want to use a system for genomic data - it's for a Master's project. How does ActorDB compare with these other systems?

kapilpipaliya

@kapilpipaliya

can i use this database in my elixir project?

randombrein

@randombrein

Hello, what's the best configuration to use with max_db_dize for different disk sizes ? Is something like following OK?

% ssd = 5TB, hdd = 10TB
{main_db_folder, "/ssd/actordb"},
{extra_db_folders, ["/hdd/actordb/1", "/hdd/2/actordb/2"]},
{max_db_dize, "5TB"}

Sergej Jurečko

@SergejJurecko

@randombrein the size should be whatever you think you're likely to hit

once you set extra folders, those must remain always active

as actors locally will be spread across them

randombrein

@randombrein

@SergejJurecko if I understand correct, can I set max_db_size to 15TB then providing extra_db_folders are always available ?

Sergej Jurečko

@SergejJurecko

db size is not split across all folders, it is size of lmdb file in every folder

Ashwin S

@ashwinsadeep

@SergejJurecko Does the thrift interface support anonymous connections?

The reason I'm asking is that, I'm planning on running ActorDB within a private network and do not need to authenticate connecting clients. If I can do away with the login step, that reduces one network round trip from my application which will make a considerable difference in my application performance.

Sergej Jurečko

@SergejJurecko

@ashwinsadeep it does not

azam

@azam

@SergejJurecko is there any form support for horizontal partitioning? use case would be: as a global service provider, i want a single table of customer data, but want to have the rows physically stored at my db node residing on the customer’s country on the column.

Sergej Jurečko

@SergejJurecko

@azam no

azam

@azam

@SergejJurecko Thanks!

randombrein

@randombrein

@SergejJurecko Hello, what's the following error stands for;

2018-12-05 04:38:27.195 [error] <0.136.0> actordb_local 478 Scheduler delay 575ms
2018-12-05 04:38:27.434 [error] <0.136.0> actordb_local 478 Scheduler delay 239ms
2018-12-05 04:38:27.639 [error] <0.136.0> actordb_local 478 Scheduler delay 204ms
...

I got multiple lines of this error after a couple of hours of migration application with heavy usage, using about #512 actordb clients. I get same error for both framed and buffered thift transport using binary protocol. Meanwhile these errors occur, client connections refused stated as;

Thrift: Thu Dec  5 04:38:27 2018 TSocket::open() error on socket (after THRIFT_POLL) <Host: localhost Port: 33306>Connection refused

Daniel Joanes

@djoanes

Hello, are there any other example use cases other than the filesystem one

I'm still not sure if my use-case works for this db

Daniel Joanes

@djoanes

If i model an actor as a customer and I have 200 million customers. How bad is it to do something like

ACTOR customer(*);
{{RESULT}} select {{curactor}}, count(1) from impression where event_ts > 12432342 and event_ts < 12432999;

Point 2. Is it possible to do:

ACTOR customer(101);
select {{curactor}} from impression where event_ts > 12432342 and event_ts < 12432999 having COUNT(1) >= 5;

Seems like SQLite doesn't like it. Must have a group by.

Daniel Joanes

@djoanes

I can group by NULL I suppose.

Daniel Joanes

@djoanes

Also, can anyone help understand what management of something like this in production would look like .

Daniel Joanes

@djoanes

@SergejJurecko ^

Preetam Bhosle

@preetb123

is this community active?

Denis Justinek

@djustinek

@SergejJurecko

usually responds

He’s one of the creators

Chintan Mishra

@ChintanMishra_twitter

I am getting following error while trying to build ActorDB. Can anyone guide me? Also, is the project abandoned?

==> bkdcore (compile)
/path/to/actordb/deps/bkdcore/src/bkdcore_mochi.erl:none: error in parse transform 'lager_transform': {badarg,
                                             [{erlang,integer_to_list,
                                               [{46,2}],
                                               [{error_info,
                                                 #{module =>
                                                    erl_erts_errors}}]},
                                              {lager_transform,make_varname,
                                               2,
                                               [{file,
                                                 "src/lager_transform.erl"},
                                                {line,271}]},
                                              {lager_transform,do_transform,
                                               5,
                                               [{file,
                                                 "src/lager_transform.erl"},
                                                {line,176}]},
                                              {lager_transform,walk_body,2,
                                               [{file,
                                                 "src/lager_transform.erl"},
                                                {line,90}]},
                                              {lager_transform,walk_clauses,
                                               2,
                                               [{file,
                                                 "src/lager_transform.erl"},
                                                {line,85}]},
                                              {lager_transform,walk_ast,2,
                                               [{file,
                                                 "src/lager_transform.erl"},
                                                {line,67}]},
                                              {compile,foldl_transform,3,
                                               [{file,"compile.erl"},
                                                {line,1145}]},
                                              {compile,fold_comp,4,
                                               [{file,"compile.erl"},
                                                {line,416}]}]}
Compiling /path/to/actordb/deps/bkdcore/src/bkdcore_mochi.erl failed:
ERROR: compile failed while processing /path/to/actordb/deps/bkdcore: rebar_abort
make: *** [Makefile:36: compile] Error 1

1 reply

I have checked out v0.10.29 tag. And I am using Fedora 33 with Erlang 24.0.1

I have also added the following lines to perform the build on Fedora

else ifeq ($(shell cat /etc/redhat-release | cut -d' ' -f1),Fedora)
  SHELLCMD = gcc c_src/cmdshell.c -lreadline -lncurses  -o priv/cmdshell

Chintan Mishra

@ChintanMishra_twitter

@SergejJurecko can you please help out

Sergej Jurečko

@SergejJurecko

It has not been tested on erlang 24

Chintan Mishra

@ChintanMishra_twitter

Which Erlang should I use?

@SergejJurecko

I am trying to build a dockerfile and send a pull request.

It will be mostly inspired from CouchDB but alpine flavored

Sergej Jurečko

@SergejJurecko

22 or 23 should be fine

greenz1

@craigZombie:matrix.org

[m]

Will do

Lubwama Abdul Muswawir

@muswawir_twitter

Hi everyone, have tested it on otp 24 and it works fine.

3 replies

Lubwama Abdul Muswawir

@muswawir_twitter

I tried using rebar3 but could not go past that error, so I used the deb package instead

Lubwama Abdul Muswawir

@muswawir_twitter

Like I said I installed on Ubuntu with otp 24 from the 0.10.29 deb package and it is working perfectly fine. Except for the erlang:time/0 in the actordb_client, I have no issue so far.

Lubwama Abdul Muswawir

@muswawir_twitter

The function is not used in the client but in thrift_reconnecting_client:188,120. Also erlang:get_stacktrace/0 in thrift_processor:102,177 was removed in otp24. Apart from the erlang client, the database works fine, actordb_console works fine.

Even the actordb_client works fine. Just a deprecation in its dependencies

Where communities thrive

People

Repo info

Activity