[galera]
wsrep_provider=/path/to/libgalera_smm.so
wsrep_cluster_address=gcomm://x.x.x.x,y.y.y.y,z.z.z.z
wsrep_slave_threads=1
binlog_format=row
default_storage_engine=InnoDB
innodb_autoinc_lock_mode=2
query_cache_size=0
--
You received this message because you are subscribed to the Google Groups "codership" group.
To unsubscribe from this group and stop receiving emails from it, send an email to codership-tea...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
On 2015-09-28 12:07, James Wang wrote:
> Sir,
> What is the solution please? Use an alternative to HAProxy?
Any middle man will double the hops - hence double the latency. Plus
HAProxy is single threaded and user space connection proxying is VERY
expensive, maybe more expensive than processing the actual query. Make
you conclusions.
Nothing beats direct connections to the server.
Hi Jay,What is the recommend value for wsrep_slave_threads please? We tuned it to 32.
By parallelism, did you mean to write some script with multi-threaded capability to import data?
Small transactions: how small is small please? Would you please elaborate by an example? For example, do a commit every 1000 iterations?
--
>> Any middle man will double the hops - hence double the latency. Plus
>> HAProxy is single threaded and user space connection proxying is VERY
>> expensive, maybe more expensive than processing the actual query. Make
>> you conclusions.
>>
>> Nothing beats direct connections to the server.
>>
>
>
> True, indeed. However, how do you achieve auto-fail-over then?
> Thanks in advance
>Client application logic, libglb, DNS load balancing - sky is the limit ;)