Date: Sun, 28 Oct 2007 13:18:40 +0100 From: Roland Smith <rsmith@xs4all.nl> To: Kris Kennaway <kris@freebsd.org> Cc: freebsd-amd64@freebsd.org Subject: Re: Experiences with RELENG_7 Message-ID: <20071028121840.GB65714@slackbox.xs4all.nl> In-Reply-To: <4723DB8F.8060407@FreeBSD.org> References: <20071027202252.GA40377@slackbox.xs4all.nl> <4723CD55.5090307@FreeBSD.org> <20071028000917.GC47576@slackbox.xs4all.nl> <4723DB8F.8060407@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
--aVD9QWMuhilNxW9f Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Sun, Oct 28, 2007 at 02:45:03AM +0200, Kris Kennaway wrote: >> After killing telak, the situation returned to normal. I restarted >> telak, and a couple of hours later it happened again. >=20 > Sounds like the issue is that telak is misbehaving. The scheduler will n= ot=20 > cause it to suddenly chew up CPU when it should be sleeping. The thing is, why didn't it happen on 6-STABLE? The current version of telak hasn't fundamentally changed in over a year. There are only two possible causes that I can think of; - gcc 4.x produces different code from gcc 3.x, and that causes the problem. - changes to the kernel caused the problem. Meanwhile I've started telak with 'nice -n 15', to see if that fixes the problem. :-) I also reniced the X server to -10. >> It also happened with cc1 when I was compiling a port. In that case the >> situation returned to normal when that particular instance was >> finished. That compilation was running in a terminal window that was >> iconified at the time. >=20 > This may be unrelated. The new version of gcc has higher memory=20 > requirements, and it is common for it to access swap during compilation= =20 > when the system is also busy running other tasks. If you see it happenin= g=20 > with cc again, make sure to check this. I didn't see any swap usage, but I'll keep an eye on it next time I'm rebuilding a big port. My desktop's CPU is usually not very busy, unless it= 's running a port compile or a big TeX or povray job. =20 >> In case I forgot to mention it, I deleted and recompiled all ports after >> the 6->7 update to make sure there was no old cruft lying around. I am >> running a custom kernel because I'm using GEOM_ELI. Do you need to see >> my kernel config? >=20 > What are you using GELI for? If it's encrypting data used by those=20 > processes that will of course add to the load. Only my /home partition is encrypted. The files that telak uses are on /usr. roland --=20 R.F.Smith http://www.xs4all.nl/~rsmith/ [plain text _non-HTML_ PGP/GnuPG encrypted/signed email much appreciated] pgp: 1A2B 477F 9970 BA3C 2914 B7CE 1277 EFB0 C321 A725 (KeyID: C321A725) --aVD9QWMuhilNxW9f Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4 (FreeBSD) iD8DBQFHJH4gEnfvsMMhpyURAvVJAJ4s5CDpj/7kc2lreupSg2jB3r6HMQCbBoTj e/VjOPTVjGhJUcNhGJ7soDU= =mLZ/ -----END PGP SIGNATURE----- --aVD9QWMuhilNxW9f--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20071028121840.GB65714>