Date: Wed, 05 Jan 2011 09:46:04 -0700 From: Warner Losh <imp@bsdimp.com> To: Ivan Voras <ivoras@FreeBSD.org> Cc: svn-src-projects@FreeBSD.org, src-committers@FreeBSD.org, Warner Losh <imp@FreeBSD.org> Subject: Re: svn commit: r216978 - in projects/graid/head/sys: cam/ata conf Message-ID: <4D24A04C.7070601@bsdimp.com> In-Reply-To: <AANLkTinjQdTmeXKhXd79dcEBYQRw98M%2B9NLz-3XEcLn-@mail.gmail.com> References: <201101042244.p04MiBf6057193@svn.freebsd.org> <AANLkTinjQdTmeXKhXd79dcEBYQRw98M%2B9NLz-3XEcLn-@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On 01/04/2011 16:37, Ivan Voras wrote: > On 4 January 2011 23:44, Warner Losh<imp@freebsd.org> wrote: >> Author: imp >> Date: Tue Jan 4 22:44:10 2011 >> New Revision: 216978 >> URL: http://svn.freebsd.org/changeset/base/216978 >> >> Log: >> Implement three types of forced errors in the ada layer for the >> purpose of testing uppper layers in the stack. >> >> kern.cam.ada.X.periodic_read_error specifies that every Nth read will >> fail. >> >> kern.cam.ada.X.force_read_error specifies that the next N reads will fail. >> >> kern.cam.ada.X.force_write_error specifies that the next N writes will fail. >> >> These are enabled with 'options ADA_TEST_FAILURE'. Otherwise, they >> aren't added to the kernel. Also, the sysctl stuff is only used for >> this feature, so move it under this option ifdef. > Maybe gnop would help you? (though on a different layer and > nondeterministic failures) Thanks for the suggestion. I wasn't aware of that, and will add it to my toolkit. These sysctls will help me force errors when I want them while testing graid behavior in a number of different scenarios. gnop will help me cope with testing how they work out all together on flaky hardware w/o needing the flaky hardware. Warner
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4D24A04C.7070601>