On Tue, 2013-02-12 at 10:29 +0100, Piotr Szymaniak wrote: > On Wed, Jan 30, 2013 at 09:50:07AM +0400, Vyacheslav Dubeyko wrote: > > Hi Piotr, > > > > On Tue, 2013-01-29 at 23:10 +0100, Piotr Szymaniak wrote: > > > Hi, > > > > > > Sorry, but I can't find an answer and noticed some checkpoints in the > > > past in the middle of other checkpoints (there was a thread, I think, > > > about similar issue with timezone change). > > > > > > Right now I have (Raspberry Pi doesn't have RTC): > > > ~ lscp > > > (...) > > > 7870 2013-01-27 15:12:30 cp - 22 42174 > > > 7871 2013-01-27 15:34:23 cp - 19 42174 > > > 7872 2013-01-27 15:42:29 cp - 22 42174 > > > 7873 1970-01-01 01:05:05 cp - 93 42176 > > > 7874 1970-01-01 01:10:05 cp - 22 42174 > > > 7875 1970-01-01 01:36:36 cp - 18 42174 > > > (...) > > > 8031 1970-01-03 07:08:19 cp - 57 42183 > > > 8032 1970-01-03 07:13:20 cp - 36 42182 > > > 8033 2013-01-29 00:01:57 cp - 30 42182 > > > 8034 2013-01-29 00:19:34 cp - 599 42313 > > > > > > I just wonder if cleaner is now smart enough to clean this mess. > > > > > > > Unfortunately, it is not. Such fix doesn't made yet. I think that it > > needs to modify nilfs-clean utility. But, currently, I am deeply inside > > fixing of issues with flush kernel thread and bad btree node. These > > issues are slightly complex. I tried several solutions but I haven't the > > proper yet. Maybe, a final solution will be a combination of several > > fixes that I tried. > > > > But fixing of issue with timezone change in my TODO list. > > Ok, can you (or someone else) refresh my memory what will hit me when cp > 7872 will be cleaned? cleanerd will stop cleaning? > Segments for processing by GC are selected on the basis of importance that defines by selection policy method. Currently, we have one method that works on the basis of condition: sui_lastmod < ss_nongc_ctime. The sui_lastmod is timestamp of the last modification. The ss_nongc_ctime is creation time of the last segment not for GC. With the best regards, Vyacheslav Dubeyko. > > Piotr Szymaniak. -- To unsubscribe from this list: send the line "unsubscribe linux-nilfs" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
