[DRBD-user] drbd 8.2.1 partially hanging when writing lots of data

Michael Nelson mikenel at gmail.com
Fri Dec 28 06:12:01 CET 2007

Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.


I've discovered (and found on the DRBD site) that drbd on top of loop is a
bad idea. I've switched to LVM and everything works now.

-mike


On 12/27/07, Michael Nelson <mikenel at gmail.com> wrote:
>
> >do NOT "force sync" if you "get stuck" during normal operation.
> > t>hat is just nonsense.
>
>
> I wouldn't do that normally, I just wanted to see if it that would change
> the behavior.
>
>
>
> > would be interessting to find out where exactly it hangs.
> > if these are still responsive to the console,
> > try to get a process list while it "hangs":
> > ps -eo pid,state,wchan:40,comm | grep -e "[d]rbd" -e D
>
>
> On the sender side:
>
>  4472 S -                                        drbd0_worker
>  4487 S -                                        drbd0_receiver
>  4488 S -                                        drbd0_asender
>  4659 D blk_congestion_wait                      dd
>
> Every once in a while it will block on pdflush. But mostly it waits on
> dd.
>
> My current dd command is:
>
> dd if=/dev/zero of=foo bs=512k count=2048
>
> Also, I am using a loop device underneath DRBD to test this, not a raw
> drive or LVM.
>
>
> > this is likely too high if you have only 512 MB RAM.
> > try with max-buffers 4000, max-epoch-size 4000 or less.
>
>
> Ok, I have adjusted it back to 4000.
>
> Thanks,
> -mike
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linbit.com/pipermail/drbd-user/attachments/20071227/b4385ff6/attachment.htm>


More information about the drbd-user mailing list