Hi, I have being try to track down top email users to sometime. When I do a network traffic check I can see there is about 15 times more email traffic pulled from mail server than sent. The problem I am trying to track down is which users are the culprit. I have enabled doevcot stats and I can do a doveadm stats dump user but I get a 0 for disk_input, disk_output, read_bytes and write_bytes. I do get mail_read_bytes numbers but I am not sure these numbers are correct. I there a top utility that works well so I can see email traffic from users mailboxes. The doveadm top does not work well. I also tried the perl scripts and that does not do a much better job. Thanks Kilburn
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Thu, 15 Oct 2015, Kilburn Abrahams wrote:> I have enabled doevcot stats and I can do a doveadm stats dump user but I > get a 0 for disk_input, disk_output, read_bytes and write_bytes. I do get > mail_read_bytes numbers but I am not sure these numbers are correct.my stats contain numbers greater than zero, except for very few users. Did you tested it with some account and waited a bit, incl. logins/logouts? - -- Steffen Kaiser -----BEGIN PGP SIGNATURE----- Version: GnuPG v1 iQEVAwUBVh9TCnz1H7kL/d9rAQIb4gf/T+sUh08end+qFaDozYpNM/grWmFYKL+S kkqR9xYOrZJePJkkkJIrGBOOzy9ie8ZQlfKkWXoyv4gu22Me6oSszzoDiaPekov7 P/RHVm/iUhKRFqINfMAbP91SlJwWK7Ht+MkMdSqGuwRoC2zovv1DjWUOsaMctJ/T c63nJL+Q7zlGJD1yfZwq9Tu69b4y19vn2uXGAi0yL4WgHdgK1JAk2R3U4JTyPjpx BjAd4txtTVHSKqvU3GVL83MsESg/3g80wMgZS9SkGmnRZv4BPd15EIgjNqo6fOLy 35hJ1ItgpqnApJETHEdUyD2f5y7GLrpoy3vLbWGRmUokd2OyV4DIcw==hz1I -----END PGP SIGNATURE-----
On 10/15/2015 06:17 PM, Steffen Kaiser wrote:> On Thu, 15 Oct 2015, Kilburn Abrahams wrote: > > > I have enabled doevcot stats and I can do a doveadm stats dump user > but I > > get a 0 for disk_input, disk_output, read_bytes and write_bytes. I > do get > > mail_read_bytes numbers but I am not sure these numbers are correct. > > my stats contain numbers greater than zero, except for very few users. > > Did you tested it with some account and waited a bit, incl. > logins/logouts? > > -- Steffen KaiserYes I waited more than 24 hrs but I would assume it would a few minutes would do it as well. I am using gentoo and dovecot 2.2.19. I get numbers for all columns except disk_input, disk_output, read_count, read_bytes, write_count and write_bytes which is 0. cat /proc/self/io rchar: 2068 wchar: 0 syscr: 6 syscw: 0 read_bytes: 0 write_bytes: 0 cancelled_write_bytes: 0 Is this a kernel setting that causes this read_bytes and write_bytes to be 0 as it happens on a test server as well. /Kilburn