On Mon, Jul 5, 2010 at 11:44 PM, Bob Jolliffe <[email protected]> wrote: > Searching back through some old mail I see something similar has > surfaced before in this bug report: > https://bugs.launchpad.net/dhis2/+bug/534567 > > Also there was an issue with one of the servers in India which also > used mysql where I had a look at the log file and found an > inexplicably large number of threads blocked in a write operation > which I now suspect is also related. > > It seems that some care needs to be taken to manage the mysql > connection pool over time. And as I suggested above I believe > (suspect!) the way to do that might be to proactively manage the > connection threads using c3p0's idleTestperiod property. Note I am no > hibernate or c3p0 expert. > > I do recall from the "old days" perhaps around 2.01 we were commonly > setting c3p0 related properties in the hibernate.properties file. I > also remember (but can't find a reference) some suggestion of dropping > this. Lars, do you remember why we don't have these parameters set in > the hibernate.properties any more? I have a funny feeling that they > are required to keep long running mysql installations alive and > kicking. >
Sorry, can't remember. Indians have been running dhis on mysql for a long time now, good if you could provide some input... > Regards > Bob > > On 5 July 2010 17:39, Orvalho Augusto <[email protected]> wrote: >> Thanks for help. >> >> Postgres I will use if I do not have a solution. >> >> I will try things on the connection pool. >> >> One very important note: >> I have used DHIS 2.01 for almost a year and never happened to have these >> errors on the same Fedora Server as described before. I have tried some >> snapshot versions and again never had this. >> I jumped from 2.01 to 2.04 and I start with MySQL troubles. >> >> Caveman >> >> >> On Mon, Jul 5, 2010 at 10:20 AM, Bob Jolliffe <[email protected]> wrote: >>> >>> Hi Orvalho >>> >>> Whereas I do agree postgres might solve your problems it does seem >>> like a drastic solution to this problem. The fact is mysql does >>> actually work and apparently work well so it should be possible to >>> solve this. >>> >>> Trying to think logically here ... you have another similar setup >>> where everything seems ok. I suppose the load could be quite >>> different on the two servers, but lets discount that for a bit. If >>> you are reusing old stale and broken connections then the only three >>> places i can think to look are: >>> (i) the jdbc driver (version compatibility). I think you checked that. >>> (ii) tcp/ip problems >>> (iii) the connection pool >>> >>> Regarding (ii) can you check the value of bind-address in >>> /etc/mysql/my.cnf (or whatever config file you are using). This >>> should be set to 127.0.0.1 unless you are connecting to the db from a >>> different host. At least this interface will always be up so you >>> shouldnt get intermittent network errors on it. >>> >>> Regarding (iii) I suggest (as above) that you look at setting up c3p0 >>> parameters to periodically test and discard stale connections. Does >>> anybody (maybe Indian team) have a good sample config? >>> >>> Regards >>> Bob >>> >>> On 4 July 2010 13:09, Orvalho Augusto <[email protected]> wrote: >>> > It will give some job because we developed a tool to convert data from >>> > one >>> > database to MySQL. >>> > >>> > It is the only thing I can try to do to solve. >>> > >>> > Caveman >>> > >>> > >>> > >>> > On Sun, Jul 4, 2010 at 8:16 AM, Knut Staring <[email protected]> wrote: >>> >> >>> >> Is postgres out of the question? >>> >> >>> >> On Jul 4, 2010 5:59 AM, "Orvalho Augusto" <[email protected]> wrote: >>> >> >>> >> I do not know what to do. I have found this: >>> >> >>> >> >>> >> http://dev.mysql.com/doc/refman/5.1/en/connector-j-usagenotes-troubleshooting.html >>> >> >>> >> And what makes me feel really bad is: >>> >> "22.3.5.3.4: I have a servlet/application that works fine for a day, >>> >> and >>> >> then stops working overnight MySQL closes connections after 8 hours of >>> >> inactivity. You either need to use a connection pool that handles stale >>> >> connections or use the "autoReconnect" parameter" >>> >> >>> >> and >>> >> "The autoReconnect facility is deprecated, and may be removed in a >>> >> future >>> >> release. " >>> >> >>> >> >>> >> So what we do? >>> >> >>> >> Caveman >>> >> >>> >> On Mon, Jun 28, 2010 at 12:59 PM, Bob Jolliffe <[email protected]> >>> >> wrote: > > Hi > > It looks... >>> >> >>> >> _______________________________________________ >>> >> Mailing list: https://launchpad.net/~dhis2-devs >>> >> Post to : [email protected] >>> >> Unsubscribe : https://launchpad.net/~dhis2-devs >>> >> More help : https://help.launchpad.net/ListHelp >>> >> >>> > >>> > >>> > >>> > >> >> >> >> > > _______________________________________________ > Mailing list: https://launchpad.net/~dhis2-devs > Post to : [email protected] > Unsubscribe : https://launchpad.net/~dhis2-devs > More help : https://help.launchpad.net/ListHelp > _______________________________________________ Mailing list: https://launchpad.net/~dhis2-devs Post to : [email protected] Unsubscribe : https://launchpad.net/~dhis2-devs More help : https://help.launchpad.net/ListHelp
