Sorry people, I missed the original message by Miguel Di Ciurcio Filho, so I'm going to comment it in Alan McKinnon's reply.
Alan McKinnon wrote: > On 11/04/2013 19:51, Miguel Di Ciurcio Filho wrote: >> Hi there, >> >> Reading the LPIC-2 V4 draft [1], I would like to ask what is the >> relevance of asking a candidate about how to patch the kernel, and >> only using the patch command. >> >> 1) The kernel can be easily configured at runtime or boot time. >> Over the years Linux has became more mature and it is possible easily >> change core components like schedulers, memory policies and HZ or >> multiplier parameters. The culture of pushing features to Linus' tree >> is a success. >> >> 2) Patching kernel source code is extremely rare and IMHO should be avoided. >> Kernels patching is indeed seldom necessary, but I do consider it a topic worthy of figuring in LPIC-2s' objectives. I often do patch kernels, as the machines I manage usually run on custom kernels, compiled from the vanilla sources. When 3.8.6 is out, I patch the 3.8.5 version with the patch-3.8.6.xz file downloaded from ftp.kernel.org, run make oldconfig and rebuild. One might wonder how relevant this is to an LPIC-2 candidate. I think it is enough to grant it a presence among the objectives. The reasons are: 1) patching and compiling are basic, fundamental skills for any developer, and developing these skills on the kernel sounds just natural to me. 2) Kernel hacking is everyday's work for people working on embedded systems, and we all know how relevant the embedded field has become. There is no embedded Linux certification right now, and I think it is still far away from being available. Kernel patching is an important topic for anyone who intends to work in the embedded sector. [...] > 3) If one _REALLY_ must maintain a kernel, one must use git. LPIC-2 is not focused with becoming a kernel maintainer. >> IMHO it is completely insane to just apply patches to a kernel without >> using git. Developers and high level SysAdmins these days cherry-pick >> features using git. There is no need of git to apply official kernel patches to the official kernel sources, or distro-provided patches to distro-provided sources. >> So, 201.3 should ask a candidate about basic git usage regarding >> kernel patching, or it must be removed, because I really don't >> remember having to patch any kernel for a long time and I strongly >> recommend to my students to not do it. > I fully agree. Kernel patching is a valuable skill, but not for the LPI > cert's target market - corporate and professional sysadmins. Where I > work, kernel patching will get you in real trouble real quick. Instead, > we must log tickets with RedHat and get a vendor-supported kernel with > features added. > > The only distro where this skill is advantageous for our purposes would > be Gentoo. This is true, but I still think a little bit of content can and should be there that covers the needs of non prime-targets of the certification. Greetings, -- Alessandro Selli Tel: 340.839.73.05 http://alessandro.route-add.net, VOIP: sip:dhatarat...@ekiga.net _______________________________________________ lpi-examdev mailing list lpi-examdev@lpi.org http://list.lpi.org/cgi-bin/mailman/listinfo/lpi-examdev
