于 2011年05月21日 00:50, Harald Hoyer 写道: > Am 20.05.2011 18:49, schrieb microcai: >> 于 2011年05月20日 23:35, Kay Sievers 写道: >>> On Fri, May 20, 2011 at 17:32, microcai<micro...@fedoraproject.org> >>> wrote: >>> >>>> Now I've full relized that, the problem is fsck-root.service. >>>> >>>> udev.service stay idle as soon as fsck-root.service exit. >>>> >>>> It is fsck-root.service that took so long to do the job. >>> >>> Check that the clock is right, that rtc is set by the kernel: >>> [ 0.851573] rtc_cmos 00:07: setting system clock to 2011-05-20 >>> 15:24:49 UTC (1305905089) >>> >>> so that fsck does not think it needs to check? >>> >>> Kay >> >> It's some kinds wired strange. >> >> When I disable readahead, fsck-root.service done in less than 1s. >> >> But when I enable readahead, fsck-root.service took more thatn 8s to >> finish. >> >> It's that strange? > > No.. it isn't... both do heavy I/O and share the same slow disk :)
No, I've change the fsck-root.service to use /sbin/fsck -p /dev/sda2 directly systemd-anaylz shows that systemd started fsck-root.service very early. But I notice fsck's ouput only at the end of the boot process. Wired strange thing! It means, fsck speed 8s more to initialize it self! > _______________________________________________ > systemd-devel mailing list > systemd-devel@lists.freedesktop.org > http://lists.freedesktop.org/mailman/listinfo/systemd-devel
signature.asc
Description: OpenPGP digital signature
_______________________________________________ systemd-devel mailing list systemd-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/systemd-devel
