Date: Thu, 30 Mar 2006 22:46:14 +0200 From: usleepless@gmail.com To: Miguel <mmiranda@123.com.sv> Cc: freebsd-questions@freebsd.org Subject: Re: terrible performance in 6.1beta4 Message-ID: <c39ec84c0603301246w2bc87b0eh57c9a601b81f35f6@mail.gmail.com> In-Reply-To: <442C3B17.4060308@123.com.sv> References: <442B2FC6.9040001@123.com.sv> <20060330011834.GA84658@xor.obsecurity.org> <c39ec84c0603300047u5530fc1fjb1ba93fcafcd490d@mail.gmail.com> <442BF0BB.8010504@123.com.sv> <20060330150136.GA12982@xor.obsecurity.org> <c39ec84c0603301151s77b37291x3ac1f79582cd753e@mail.gmail.com> <442C3B17.4060308@123.com.sv>
next in thread | previous in thread | raw e-mail | index | archive | help
Miguel, > > > Yes, it is a dump of a single table. i want to tranfer the data from one= =20 > server to another, and this is one of the biggest table. ok, but gentoo performs the same task ok, so it is not a postgresql problem= . you have not confirmed wether the gentoo-box is running with the same pos= tgresql.conf. is it? if not, which are the differences? > attached is my config file, shared_buffers are 25% of total RAM i my experience, you should go much higher. if the server is not in product= ion yet, you might go as high as 75% ( assuming no other processes need th= ese kind of resources ). i am not familiar with the work_mem flag, it was = not there last time i tuned a postgresql. you might consider upping your wa= l-buffers, but i am not sure if they are used by copyin. > im guessing that this is a disk controlled bug or something, when i=20 > execute any query involving many rows, the server response is very low,= =20 > ssh, su, even copy or rename a file, cpu usage remains ~87% idle though ofcourse it might be hardware related: have you checked your diskperformanc= e? what is your throughput? how does this throughput compare with the gento= o box? regards, usleep > --- > Miguel > miguel >=20 >=20
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?c39ec84c0603301246w2bc87b0eh57c9a601b81f35f6>