Hello again!
I understand that it's not recommended running osd and rbd-nbd on the same host and i actually moved my rbd-nbd to a completely clean host (same kernel and OS though), but with same result.
I hope someone can resolve this and you seem to indicate it is some kind of known error but i didn't really understand the github commit that you linked.
If other logs or info is needed i'm happy to provide it.
//Stefan
________________________________________
Från: Ilya Dryomov [idryomov@xxxxxxxxx]
Skickat: den 25 april 2016 17:31
Till: Stefan Lissmats; Mykola Golub
Kopia: Mika c; ceph-users
Ämne: Re: RBD image mounted by command "rbd-nbd" the status is read-only.
On Mon, Apr 25, 2016 at 1:53 PM, Stefan Lissmats <stefan@xxxxxxxxxx> wrote:
> Hello!
>
> Running a completely new testcluster with status HEALTH_OK i get the same
> error.
> I'm running Ubuntu 14.04 with kernel 3.16.0-70-generic and ceph 10.2.0 on
> all hosts.
> The rbd-nbd mapping was done on the same host having one osd and mon. (This
> is a small cluster with 4 virtual hosts and one osd per host).
>
> Steps after creating cluster.
>
> Created rbd device with standard options.
> #rbd create --size 50G nbd2
>
> Map the device (became device /dev/nbd2)
> #rbd-mbd map nbd2
>
> Create ext4 filesystem
> #mkfs.ext4 /dev/nbd2
>
> During creation of filesystem there was alot of errors in dmesg but mkfs
> indicated done.
> The errors was block nbd2: Other side returned error (5)
>
> I was able to mount the ext4 filesystem but that created even more errors in
> dmesg.
>
> Here is a selection of dmesg that probably contains the intresting bits.
>
> [13864.102569] block nbd2: Other side returned error (5)
> [13951.186296] block nbd2: Other side returned error (5)
> [13951.186443] blk_update_request: 2157 callbacks suppressed
> [13951.186445] end_request: I/O error, dev nbd2, sector 0
> [13951.186598] quiet_error: 271152 callbacks suppressed
> [13951.186600] Buffer I/O error on device nbd2, logical block 0
> [13951.186780] lost page write due to I/O error on nbd2
> [13951.187816] EXT4-fs (nbd2): mounted filesystem with ordered data mode.
> Opts: (null)
> [13952.049103] block nbd2: Other side returned error (5)
> [13952.049323] end_request: I/O error, dev nbd2, sector 8464
> [13952.070722] block nbd2: Other side returned error (5)
> [13952.071009] end_request: I/O error, dev nbd2, sector 8720
> [13952.074069] block nbd2: Other side returned error (5)
> [13952.074392] end_request: I/O error, dev nbd2, sector 8976
> [13952.075283] block nbd2: Other side returned error (5)
> [13952.075635] end_request: I/O error, dev nbd2, sector 9232
> [13952.076249] block nbd2: Other side returned error (5)
> [13952.076636] end_request: I/O error, dev nbd2, sector 9488
> [13952.077108] block nbd2: Other side returned error (5)
> [13952.077606] end_request: I/O error, dev nbd2, sector 9744
> [13952.078064] block nbd2: Other side returned error (5)
> [13952.078537] end_request: I/O error, dev nbd2, sector 10000
> [13952.079038] block nbd2: Other side returned error (5)
> [13952.079583] end_request: I/O error, dev nbd2, sector 10256
> [13952.080301] block nbd2: Other side returned error (5)
> [13952.080869] end_request: I/O error, dev nbd2, sector 10512
> [13952.081474] block nbd2: Other side returned error (5)
> [13952.082088] block nbd2: Other side returned error (5)
> [13952.082701] block nbd2: Other side returned error (5)
> [13952.083316] block nbd2: Other side returned error (5)
> [13952.083943] block nbd2: Other side returned error (5)
> [13952.084654] block nbd2: Other side returned error (5)
> [13952.085301] block nbd2: Other side returned error (5)
Looks like this has come up before:
https://github.com/ceph/ceph/pull/7215/commits/3ff60a61bf68516983c0b6ea6791ce712c98a073
Do we set rval to the length of the request for aio writes? I thought
we did this only for reads and that it's always <= 0 on writes.
Mykola, could you look into this?
I certainly wouldn't advise running rbd-nbd on OSD hosts.
Thanks,
Ilya
_______________________________________________
ceph-users mailing list
ceph-users@xxxxxxxxxxxxxx
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
