Date: Wed, 22 Jan 2003 20:43:45 +0200 From: Maxim Sobolev <sobomax@portaone.com> To: Edwin Groothuis <edwin@mavetju.org> Cc: portmgr@freebsd.org, ports@freebsd.org Subject: Re: Bento Reporting Enhancements Message-ID: <3E2EE661.71883F63@portaone.com> References: <20030120132715.GA27491@k7.mavetju>
next in thread | previous in thread | raw e-mail | index | archive | help
Edwin Groothuis wrote: > > Hi, > > I was wondering if Port Building Reporting could be done better. > Right now, the building of ports is done on several different > host-systems, different OSses and at the end you have a log-file > which gets batch-processed into a couple of HTML files. There is > no way to quickly check the history of the build of a port (when > did it start to stop building correctly, different per platform and > OS version of course (what did you expect :-), you can't quickly > see if the port builds, or doesn't, on a different platform and > there is no way to get a small overview of all ports failing with > a perl-error, you get the information of all ports. Etc etc etc. > > Yes, to do these things you need a database. Well, let us get a > database then. If after each build (or attemtp to build :-) the > logfile and some meta-information (which processor, which OS version) > is send to an account@machine which processes it and stores it in > the database. Then with a webbrowser, webserver and a couple of PHP > scripts you can access the data there. Rocket-science, isn't it? > > With this, you have a couple of advantages: > > - Seperation of the port-building and the reporting. > - History of building. (failures, successes, versions) > - More information in the reports (information about all platforms, OSs etc) > - "Optimized and personal" reporting, no more huge chunks of > pre-processed text. > - Easier post-processing of the data (think monthly reports, think > email-reporting to the maintainers, commitors) > > With this, you also have disadvantages: > > - You have to run a database, which costs CPU+Disk+RAM. > - The website becomes dynamic, which costs CPU. > > I've been told where I can find some old logfiles and am willing > to spend some time to setup a prototype for it. I've seen other > people (hi Mike) who I can ask to help me with it. > If I can make such a beast, would there be a chance that this would > become an official service on the FreeBSD server-farm? Or are the > disadvantages too big? In general, I like the proposal, but there are several technical and non-technical difficulties: 1. Full set of logs (both positive and negative ones) from one build is quite large, therefore there might be a problem keeping some more of less long history in DB. 2. I am not sure that it will be possible to allocate machine for serving database/php/web. Not only availability of the actual free hardware, but also availability of rack space in the yahoo's datacenter matters. 3. Not being a portmgr@ member you are unlikely to be given any administrative access necessary for managing the database to a machine in the cluster for security reasons. For example, even not all portmgr@ members have root shell on bento cluster and it took several months to get even such restricted access. Therefore, even if (1) and (2) resolved, somebody with such access will have to manage database for you. -Maxim To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-ports" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3E2EE661.71883F63>