looks like http://tracker.ceph.com/issues/18314. please try: run "ceph fs set cephfs1 allow_dirfrags 1” and set mds config mds_bal_frag to 1, set mds_bal_split_size to 5000, and set mds_bal_fragment_size_max 50000. > On 11 Sep 2017, at 15:46, donglifecomm@xxxxxxxxx wrote: > > ZhengYan, > > kernel client is 4.12.0. > > [root@yj43959-ceph-dev ~]# uname -a > Linux yj43959-ceph-dev.novalocal 4.12.0-1.el7.elrepo.x86_64 #1 SMP Sun Jul 2 20:38:48 EDT 2017 x86_64 x86_64 x86_64 GNU/Linux > > Thanks a lot. > > donglifecomm@xxxxxxxxx > > From: Yan, Zheng > Date: 2017-09-11 15:24 > To: donglifecomm@xxxxxxxxx > CC: ceph-users > Subject: Re: cephfs(Kraken 11.2.1), Unable to write more file when one dir more than 100000 files, mds_bal_fragment_size_max = 5000000 > > > On 11 Sep 2017, at 14:07, donglifecomm@xxxxxxxxx wrote: > > > > ZhengYan, > > > > I set "mds_bal_fragment_size_max = 100000, mds_bal_frag = true", then I write 100000 files named 512k.file$i, but there are still some file is missing. such as : > > [root@yj43959-ceph-dev cephfs]# find ./volumes/ -type f | wc -l > > 91070 > > which client do you use? kernel client or fuse client and version? > > Yan, Zheng > > > > > <InsertPic_.png> > > > > donglifecomm@xxxxxxxxx > > > > From: Yan, Zheng > > Date: 2017-09-08 15:20 > > To: donglifecomm@xxxxxxxxx > > CC: ceph-users; marcus.haarmann > > Subject: Re: cephfs(Kraken 11.2.1), Unable to write more file when one dir more than 100000 files, mds_bal_fragment_size_max = 5000000 > > > > > On 8 Sep 2017, at 13:54, donglifecomm@xxxxxxxxx wrote: > > > > > > ZhengYan, > > > > > > I'm sorry, just a description of some questions. > > > > > > when one dir more than 100000 files, I can continue to write it , but I don't find file which wrote in the past. for example: > > > 1. I write 100000 files named 512k.file$i > > > > > > 2. I continue to write 10000 files named aaa.file$i > > > > > > 3. I continue to write 10000 files named bbb.file$i > > > > > > 4. I continue to write 10000 files named ccc.file$i > > > > > > 5. I continue to write 10000 files named ddd.file$i > > > > > > 6. I can't find all ddd.file$i, some ddd.file$i missing. such as: > > > > > > [root@yj43959-ceph-dev scripts]# find /mnt/cephfs/volumes -type f | grep 512k.file | wc -l > > > 100000 > > > [root@yj43959-ceph-dev scripts]# ls /mnt/cephfs/volumes/aaa.file* | wc -l > > > 10000 > > > [root@yj43959-ceph-dev scripts]# ls /mnt/cephfs/volumes/bbb.file* | wc -l > > > 10000 > > > [root@yj43959-ceph-dev scripts]# ls /mnt/cephfs/volumes/ccc.file* | wc -l > > > 10000 > > > [root@yj43959-ceph-dev scripts]# ls /mnt/cephfs/volumes/ddd.file* | wc -l // some files missing > > > 1072 > > > > It’s likely caused by http://tracker.ceph.com/issues/18314. To support very large directory, you should enable directory fragment instead of enlarge mds_bal_fragment_size_max. > > > > Regards > > Yan, Zheng > > > > > > > > <InsertPic_.png> > > > > > > donglifecomm@xxxxxxxxx > > > > > > From: donglifecomm@xxxxxxxxx > > > Date: 2017-09-08 13:30 > > > To: zyan > > > CC: ceph-users > > > Subject: cephfs(Kraken 11.2.1), Unable to write more file when one dir more than 100000 files, mds_bal_fragment_size_max = 5000000 > > > ZhengYan, > > > > > > I test cephfs(Kraken 11.2.1), I can't write more files when one dir more than 100000 files, I have already set up "mds_bal_fragment_size_max = 5000000". > > > > > > why is this case? Is it a bug? > > > > > > Thanks a lot. > > > > > > donglifecomm@xxxxxxxxx _______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
 |