Hi,

I have retry the test, but with FUSE CephFS client. It seem everything is
OK. Any explainantion? Is Kernel CephFS client less featured (more limited)
and/or less stable than FUSE CephFS client like on RBD?

Best regards,


> Date: Thu, 25 Aug 2016 00:25:22 +0700
> From: Lazuardi Nasution <mrxlazuar...@gmail.com>
> To: ceph-users@lists.ceph.com
> Subject: [ceph-users] CephFS Big Size File Problem
> Message-ID:
>         <CA+u3GuLMxEt4hhyixcG45PXritGyt9N_t73aiHtb1XQS5OPm-g@mail.gmail.
> com>
> Content-Type: text/plain; charset="utf-8"
>
> Hi,
>
> I have problem with CephFS on writing big size file. I have found that my
> OpenStack Nova backup was not working after I change the rbd based mount of
> /var/lib/nova/instances/snapshots to CephFS based (mounted on /etc/fstab
> on
> all Nova compute nodes). I couldn't relize the cause until I tried to do
> some test below.
>
> [root@compute-aa instances]# sudo -S -u nova dd if=/dev/zero
> of=./snapshots/test/test.dat bs=4096 count=1024
> 1024+0 records in
> 1024+0 records out
> 4194304 bytes (4.2 MB) copied, 0.002956 s, 1.4 GB/s
>
> It seem that making 4.2MB size file is OK. The "test" folder was been
> created by nova user too.
>
> [root@compute-aa instances]# sudo -S -u nova dd if=/dev/zero
> of=./snapshots/test/test.dat bs=4096 count=1048576
>
> This test was hang and gave me "D state" of dd process which cannot be
> killed, so I must reboot that node. This make afraid to do some more test
> with lower size. When I tried to see the folder content, I found following
> result.
>
> [root@compute-aa ~]# ls -lah /var/lib/nova/instances/snapshots/test/
> total 0
> drwxr-xr-x 1 nova nova 0 Aug 24 22:35 .
> drwxr-xr-x 1 nova nova 0 Aug 24 22:33 ..
> -rw-r--r-- 1 nova nova 0 Aug 24 22:35 test.dat
>
> I prefer to use CephFS since the snapshot folder is only for temporary.
> Please help me to solve this problem.
>
> Best regards,
>
_______________________________________________
ceph-users mailing list
ceph-users@lists.ceph.com
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com

Reply via email to