Am Montag, 15. September 2014, 13:12:35 schrieb Martin Steigerwald: > Am Montag, 15. September 2014, 11:59:17 schrieb Martin Steigerwald: > > Am Montag, 15. September 2014, 11:20:38 schrieb Martin Steigerwald: > > > Am Mittwoch, 10. September 2014, 23:25:21 schrieb Zhang Cheng: > > > > Hi there, > > > > > > Hello Cheng, hello Marc, > > > > > > > I have made package for version 2.0.2, and uploaded to mentors: > > > > http://mentors.debian.net/package/atop > > > > > > Thank you for your work. I certainly want to see more recent atop for > > > Jessie. > > > > > > > Made two commits on top of Vcs-Git: > > > > * Import upstream 2.0.2 code > > > > > > > > https://github.com/StephenPCG/debian-atop/commit/527db3149be62b097b7ed > > > > cf > > > > 70 > > > > 2a e32825aa0eb50 * refresh patches to fit new code, remove > > > > manpage-syntax > > > > patch (applied upstream) > > > > > > > > https://github.com/StephenPCG/debian-atop/commit/bc0b6f499a8c947bf4506 > > > > 25 > > > > 72 > > > > 19 1d9656d033a7d > > > > > > > > Is there anyone who can sponsor me? > > > > > > Cheng, after coordinating with Marc, the maintainer, I made a 2.1 > > > package > > > already, but I am not completely satisfied since I had strange merge > > > conflicts on git-import-orig of the 2.1 archive in top of the 2.0.2 work > > > Marc already did. I contacted Marc about it, but didn't yet get a reply. > > > > > > I have it on github at the moment: > > > > > > https://github.com/teamix/atop-debian > > > > > > So I am planning to redo this from scratch and import 2.1 directly on > > > top > > > of the latest git alioth version 1.2.7-something today. > > > > > > So I branched the current situation into martin-first-try and will redo > > > from scratch today. > > > > Okay, I got a clean 2.1 git-import-orig on > > > > https://github.com/teamix/atop-debian/ > > > > now. So from 1.27 version the git-import-orig worked without issues… > > > > I will now redo patches and other fixes I Marc and I made for the 2.0.2 > > and > > previous 2.1 attempt and work from there. > > > > I hope I have a workable package in a hour or so… well probably after > > midday break. > > Okay, this works now and is much cleaner than my previous attempt. Marc I > redid your changes. > > I will do more changes and lintian fixes, try to fix up atopacctd not being > started and look into systemd stuff after midday break. > > All on github in master branch.
Okay, I am a lot further. I made a first attempt at debianizing the atopacct initscript. I also fixed the init script of atop not to add daily.log and atop_somedate.log as arguments, producing a syntax error. For now I use the atop_somedate thing as atopsar expects it. Here is a (probably incomplete) list of issues TODOs I still have: === how to deal with incompatible atop log file format on upgrades? === atop doesn't start: + start-stop-daemon --start --background --quiet --pidfile /var/run/atop.pid -- make-pidfile --startas /usr/share/atop/atop.wrapper -- /usr/bin/atop -a -w /var/log/atop/atop_20140915 600 + case "$?" in + '[' no '!=' no ']' + exit 0 mango:[…]# /usr/bin/atop -a -w /var/log/atop/atop_20140915 600 existing file /var/log/atop/atop_20140915 has incompatible header (created by version 2.0 - current version 2.1) atopsar does not like it either: mango:[…]# atopsar raw file /var/log/atop/atop_20140915 has incompatible format (created by version 2.0 - current version 2.1) trying to activate atopsar-2.0.... activation of atopsar-2.0 failed! This probably needs a debconf to ask to remove the old ones? Well after a day it issue should not happen by default anymore, but still. === systemd integration === service files, then I probably could also undo the exit 0 fix in atopacct file for the case its running, systemd doesn't like that. === Review === And I also need a review of the atopacct init script, reload is not implemented yet, and this is my first init script hacking in ages. start, stop, status, force-reload appear to work and is lintian clean, but I don't like that it doesn't work with pid file at the moment. atopacctd double forks itself away from any session: 254 /* 255 ** daemonize this process 256 ** i.e. be sure that the daemon is no session leader (any more) 257 ** and get rid of a possible bad context that might have been 258 ** inherited from ancestors 259 */ 260 if ( fork() ) // implicitly switch to background 261 exit(0); // continue in child process 262 263 setsid(); // become session leader to lose ctty 264 265 if ( fork() ) // continue in child process 266 exit(0); // --> no session leader, no ctty === Extra goodie === packages for netatop and netatop-dkms :) Won't promise on that tough. Cheng, interested? Done for today I think. Package is quite lintian happy already: lintian -IE atop_2.1-1_amd64.changes I: atop source: debian-watch-contains-dh_make-template I: atop: hyphen-used-as-minus-sign usr/share/man/man1/atop.1.gz:1238 I: atop: hyphen-used-as-minus-sign usr/share/man/man1/atop.1.gz:1291 I: atop: hyphen-used-as-minus-sign usr/share/man/man1/atop.1.gz:1458 I: atop: hyphen-used-as-minus-sign usr/share/man/man8/atopacctd.8.gz:8 Ciao, Martin -- Martin Steigerwald | Consultant / Trainer teamix GmbH Südwestpark 43 90449 Nürnberg Tel.: +49 911 30999 55 | Fax: +49 911 30999 99 mail: [email protected] | web: http://www.teamix.de | blog: http://blog.teamix.de Amtsgericht Nürnberg, HRB 18320 | Geschäftsführer: Oliver Kügow, Richard Müller **Unsere Networking Veranstaltungen im September - JETZT ANMELDEN** N-IX Stammtisch - 16.09.2014 - Anmeldung unter: http://www.teamix.de/go/N-IX-Stammtisch/ Healthcare IT Stammtisch - 23.09.2014 - Anmeldung unter: http://www.teamix.de/go/Healthcare-IT-Stammtisch/ -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected]
