[pmacct-discussion] nfacctd: mysql Plugin -- DB Writer [total]

2017-03-02 Thread Luc Perreau
Hi All, I am having problems with my mysql db staying in the UPDATE stage even if i set sql_locking_style: row or not. My processes reach 10 as shown here: Feb 27 05:59:01 WARN ( total/mysql ): Maximum number of SQL writer processes reached (10). Feb 27 06:00:01 WARN ( total/mysql ): Maximum num

Re: [pmacct-discussion] aggregate_filter for non pcap elements

2017-03-02 Thread Paolo Lucente
Hi Timo, Thanks for providing the use-cases, they sound valid to me. Also, you are right saying it's no big deal to implement these new keys as they would be mostly similar to the existing ones. I would make one exception which did not raise from your email: the potential need for IP addresses (i

Re: [pmacct-discussion] bgp_daemon wrong next-hop

2017-03-02 Thread Paolo Lucente
Mostly for the archives: the following patch seems to have worked for the BGP next-hop issue Catalin was experiencing. Waiting for feedback from Alberto all looks good from him too: https://github.com/pmacct/pmacct/commit/89047c047fddfe2ec4773ddd9f7e18522af35e26 Cheers, Paolo On Tue, Feb 28, 20

Re: [pmacct-discussion] nfacctd: mysql Plugin -- DB Writer [total]

2017-03-02 Thread Paolo Lucente
Hi Luc, At a glance it looks a case where you are overwhelming the RDBMS 1) you write data to a static table, ie. acct, likely making the table and its index(es) big; 2) you make use of UPDATE queries, which are expensive; whereas you should try to aim at an INSERT-only environment (*); 3) the ag

Re: [pmacct-discussion] nfacctd: mysql Plugin -- DB Writer [total]

2017-03-02 Thread Luc Perreau
Thanks Paolo, I'll see how i can enhance this setup. Kind regards, Luc On Fri, Mar 3, 2017 at 8:23 AM, Paolo Lucente wrote: > > Hi Luc, > > At a glance it looks a case where you are overwhelming the RDBMS 1) you > write data to a static table, ie. acct, likely making the table and its > index