Thank you very much to all team for fast deploy of these packages. I see now that galera + mariadb-galera is on sabayon-weekly repository but I see that on package galera garbd use flag is missing. Could be helpful add Galera Arbitrator daemon on sys-cluster/galera package.
Thanks again. G. 2015-10-24 22:55 GMT+02:00 Joost Ruis <[email protected]>: > 1, 2 and 4 are done. > > On Sat, Oct 24, 2015 at 6:34 PM, Ettore Di Giacinto < > [email protected]> wrote: > >> Hi Geaaru, >> >> First, thank you for the suggestions, >> >> answering in order: >> >> 1) noted, soon will hit entropy >> >> 2) done >> https://github.com/Sabayon/build/commit/f37805526ef0d7cbfacc7764a7ee8b5e3e826ca2 >> , >> in Entropy soon >> >> 3) the eselect would be a great approach, feel free to do it and open a >> pull request, i'll be happy to review it. Have you tried to blacklist the >> module? zfs is actually handy, dunno if worth deleting the file from the >> packages, this usually means that we must handle then both packages >> manually (in reality, we are already doing that for zfs, but not for xen). >> I try to avoid forks, just when strictly necessary (because requires more >> maintenance) >> >> 4) >> https://github.com/Sabayon/build/commit/1f40f115dde095d97fd38dc1ceec9d4276485a9b >> , >> in Entropy soon >> >> 4-bis?) Never experienced that, file a bug with logs /steps to reproduce >> >> 5) noted, but since is not a priority, you could speed up things opening >> a pull request for it >> >> 6) Fabio takes care of the Sabayon kernel, IMO those options sounds good >> to me. >> >> Cheers, >> Ettore >> >> >> >> >> Il giorno ven 23 ott 2015 alle ore 10:49 Geaaru <[email protected]> ha >> scritto: >> >>> Hi guys, >>> >>> hereinafter some useful packages and/or use flag to apply on sabayon >>> server release and on sabayon desktop release for improve : >>> >>> 1) currently sabayon release doesn't provide package about mariadb >>> -galera (an enhancement of mariadb database with support to galera >>> plugin that permit manage of cluster of database with innodb engine). I >>> think that currently mariadb-galera is the best cluster solution for >>> Mysql database and it is an error doesn't provide package about this on >>> sabayon repository above all on sabayon server release. >>> >>> 2) Module app-emulation/qemu doesn't include these use flags: >>> >>> +glusterfs (for add support on directly use glusterfs library for integ >>> ration to glusterfs filesystem. Else it use fuse feature with bad >>> performance) >>> >>> +nfs (add NFS support) >>> >>> +rbd (add support to ceph block filesystem) >>> >>> On openstack platform currenctly I use qemu with gluster and nfs. >>> >>> 3) on sabayon release automatically zfs module is installed inside >>> /usr/lib/modules-load.d/zfs.conf. I think that this is not useful on >>> sabayon-desktop where often zfs filesystem is not used and also inside >>> sabayon vm. Also, I think that could be correct avoid install of >>> zfs.conf file inside /usr/lib/modules-load/ directory on every upgrade >>> of the syste. Every time I upgrade sabayon vm I need remember to remove >>> or rename zfs.conf to avoid load of zfs kernel module. Same thing for >>> xen modules. >>> Best solution is handle these files under modules-load.d directory >>> through eselect but this require a develop of eselect extension and >>> related ebuild. As first solution could be fine remove automatic >>> install of zfs.conf file. >>> >>> 4) Module app-emulation/libvirt doesn't include these use flags: >>> +glusterfs (support of gluster filesystem) >>> >>> +lxc (support to linux container) >>> >>> +rbd (support to ceph block filesystem) >>> >>> +virtualbox (support to virtualbox) >>> >>> +fuse (support to fuse... this with less priority) >>> >>> I use these flags with openstack platform and on sabayon-server release >>> colud >>> >>> 4) Avoid automatic activation of cronyd daemon. On my sabayon vm I >>> configured systemd-timesyncd or ntpd daemon for time sync but on >>> upgrade vm I see that cronyd is been automatically activated as a >>> service. This avoid then that previous configuration of systemd >>> -timesyncd and ntpd works correctly. Do you think that is correct >>> activate automatically this service on upgrade of the system ? >>> >>> 5) Journald configuration: Default installation of systemd (from gentoo >>> portage) doesn't provide any smart configuration of systemd-journald >>> service because doesn't configure systemd-journald to handle log rotate >>> or something like this. So, if a user doesn't reconfigure this file >>> after some days will optain a disk full. This, is a problem on ebuild >>> but fwik systemd ebuild is been overwrite from sabayon release, so >>> could be fine correct this file at least >>> >>> 6) I could be fine add kernel options: >>> - ipset modules (I currently use ipset with iptables) >>> - nfsv4.1 support >>> - enable CONFIG_NET_CLS_CGROUP >>> >>> For now it is all. >>> Thanks >>> >>> My cent. >>> >>> G. >>> >>> >>> >> >> >> > > > > -- -- -- Geaaru
