8000 readme · m99coder/postgres_cluster@96e2407 · GitHub
[go: up one dir, main page]
Skip to content

Commit 96e2407

Browse files
kelvichkelvich
committed
readme
1 parent 2cb2913 commit 96e2407

File tree

1 file changed

+30
-61
lines changed

1 file changed

+30
-61
lines changed

‎README.md

Lines changed: 30 additions & 61 deletions
Original file line numberDiff line numberDiff line change
@@ -10,74 +10,43 @@ This is mirror of postgres repo with several changes to the core and few extra e
1010

1111
* Transaction manager interface (eXtensible Transaction Manager, xtm). Generic interface to plug distributed transaction engines. More info at [[https://wiki.postgresql.org/wiki/DTM]] and [[http://www.postgresql.org/message-id/flat/F2766B97-555D-424F-B29F-E0CA0F6D1D74@postgrespro.ru]].
1212
* Distributed deadlock detection API.
13-
* Fast 2pc patch. More info at [[http://www.postgresql.org/message-id/flat/74355FCF-AADC-4E51-850B-47AF59E0B215@postgrespro.ru]]
13+
* Logical decoding of two-phase transactions.
14+
1415

1516
## New extensions:
1617

17-
* pg_dtm. Transaction management by interaction with standalone coordinator (Arbiter or dtmd). [[https://wiki.postgresql.org/wiki/DTM#DTM_approach]]
1818
* pg_tsdtm. Coordinator-less transaction management by tracking commit timestamps.
19 10000 19
* multimaster. Synchronous multi-master replication based on logical_decoding and pg_dtm.
2020

2121

2222
## Changed extension:
2323

24-
* postgres_fdw. Added support of pg_dtm.
25-
26-
## Deploying
27-
28-
For deploy and test postgres over a cluster we use ansible. In each extension directory one can find test subdirectory where we are storing tests and deploy scripts.
29-
30-
31-
### Running tests on local cluster
32-
33-
To use it one need ansible hosts file with following groups:
34-
35-
farms/cluster.example:
36-
```
37-
[clients] # benchmark will start simultaneously on that nodes
38-
server0.example.com
39-
[nodes] # all that nodes will run postgres, dtmd/master will be deployed to first
40-
server1.example.com
41-
server2.example.com
42-
server3.example.com
43-
```
44-
45-
After you have proper hosts file you can deploy all stuff to servers:
46-
47-
```shell
48-
# cd pg_dtm/tests
49-
# ansible-playbook -i farms/sai deploy_layouts/cluster.yml
50-
```
51-
52-
To perform dtmbench run:
53-
54-
```shell
55-
# ansible-playbook -i farms/sai perf.yml -e nnodes=3 -e nconns=100
56-
```
57-
58-
here nnodes is number of nudes that will be used for that test, nconns is the
59-
number of connections to the backend.
60-
61-
62-
63-
## Running tests on Amazon ec2
64-
65-
66-
In the case of amazon cloud there is no need in specific hosts file. Instead of it
67-
we use script farms/ec2.py to get current instances running on you account. To use
68-
that script you need to specify you account key and access_key in ~/.boto.cfg (or in
69-
any other place that described at http://boto.cloudhackers.com/en/latest/boto_config_tut.html)
24+
* postgres_fdw. Added support of pg_tsdtm.
25+
26+
## Installing multimaster
27+
28+
1. Build and install postgres from this repo on all machines in cluster.
29+
1. Install contrib/raftable and contrib/mmts extensions.
30+
1. Right now we need clean postgres installation to spin up multimaster cluster.
31+
1. Create required database inside postgres before enabling multimaster extension.
32+
1. We are requiring following postgres configuration:
33+
1. 'max_prepared_transactions' > 0 -- in multimaster all writing transaction along with ddl are wrapped as two-phase transaction, so this number will limit maximum number of writing transactions in this cluster node.
34+
1. 'synchronous_commit - off' -- right now we do not support async commit. (one can enable it, but that will not bring desired effect)
35+
1. 'wal_level = logical' -- multimaster built on top of logical replication so this is mandatory.
36+
1. 'max_wal_senders' -- this should be at least number of nodes - 1
37+
1. 'max_replication_slots' -- this should be at least number of nodes - 1
38+
1. 'max_worker_processes' -- at least 2*N + P - 1, where N is number of nodes in cluster, P size of pool of workers(see below) (1 raftable, n-1 receiver, n-1 sender, mtm-sender, mtm-receiver, + number of pool worker).
39+
1. 'default_transaction_isolation = 'repeatable read'' -- multimaster isn't supporting default read commited level.
40+
1. Also multimaster have following configuration parameters:
41+
1. 'multimaster.conn_strings' -- connstrings for all nodes in cluster, separated by comma.
42+
1. 'multimaster.node_id' -- id of current node, number starting from one.
43+
1. 'multimaster.workers' -- number of workers that can apply transactions from neighbouring nodes.
44+
1. 'multimaster.use_raftable = true' -- just set this to true. Deprecated.
45+
1. 'multimaster.queue_size = 52857600' -- queue size for applying transactions from neighbouring nodes.
46+
1. 'multimaster.ignore_tables_without_pk = 1' -- do not replicate tables without primary key
47+
1. 'multimaster.heartbeat_send_timeout = 250' -- heartbeat period (ms).
48+
1. 'multimaster.heartbeat_recv_timeout = 1000' -- disconnect node if we miss heartbeats all that time (ms).
49+
1. 'multimaster.twopc_min_timeout = 40000' -- rollback stalled transaction after this period (ms).
50+
1. 'raftable.id' -- id of current node, number starting from one.
51+
1. 'raftable.peers' -- id of current node, number starting from one.
7052

71-
To create VMs in cloud run:
72-
```shell
73-
# ansible-playbook -i farms/ec2.py deploy_layouts/ec2.yml
74-
```
75-
After that you should wait few minutes to have info about that instances in Amazon API. After
76-
that you can deploy postgres as usual:
77-
```shell
78-
# ansible-playbook -i farms/ec2.py deploy_layouts/cluster-ec2.yml
79-
```
80-
And to run a benchmark:
81-
```shell
82-
# ansible-playbook -i farms/sai perf-ec2.yml -e nnodes=3 -e nconns=100
83-
```

0 commit comments

Comments
 (0)
0