Re: BundleDBPersistenceManager and autoCommit mode

[Przemo Pakulski <pr...@cognifide.com>]

What do you mean by clustered env. exactly : Jackrabbit cluster, 
database cluster, other ? Btw, simple db persistence manager don't use 
autocommit mode, does it mean that it will not work in clustered env ?

Myabe we should add parameter to BundleDB PMs, and use autoCommit mode 
only in 'clustered environment' to avoid extra overhead in simple 
deployment.

Przemo

>> Is there any specific reason why BundlePM keep open connection in
>> autoCommit mode, then switch on/off this flag during every call of store
>> method, instead of keeping connection with autoCommit disabled.
> 
> hi,
> the reason is that in a clustered environment, the 'select' statements
> must be committed as well,  otherwise the tables remain locked. and
> instead of explicitly commit each time after a read, we used the
> autocommit as default and switch it off during store.
> 
> regards, toby

Re: BundleDBPersistenceManager and autoCommit mode

[Przemo Pakulski <pr...@cognifide.com>]

It's hard to say exact numbers, there are to many variables (network 
speed/latency, db server type, jdbc driver, change log size).
For sure it means 2 extra network calls, which could be easily avoided.

See attached example screen from JProfiler using MSSQL server, and small 
change logs, overhead is about 20%.

Cheers, Przemo

> Hi,
> 
>> What do you mean by clustered env.
> 
> Jackrabbit cluster
> 
>> use autoCommit mode only in 'clustered environment'
>> to avoid extra overhead in simple
>> deployment.
> 
> Did you test how much overhead it is? Do you have a test case that
> shows that performance can be improved by changing it? If yes, how
> much is the difference, and could you post your test case?

Re: BundleDBPersistenceManager and autoCommit mode

[Thomas Mueller <th...@gmail.com>]

Hi,

> What do you mean by clustered env.

Jackrabbit cluster

> use autoCommit mode only in 'clustered environment'
> to avoid extra overhead in simple
> deployment.

Did you test how much overhead it is? Do you have a test case that
shows that performance can be improved by changing it? If yes, how
much is the difference, and could you post your test case?

Regards,
Thomas