[DRBD-user] How to costumize the size of a DRBD partition ?

[Fabrice Durand]

Hi all,
Here is a certainly trivial question :
I am using a DRBD (v 0.7.10) partition on a machine, to make FTP
transfers or simply to copy files on it. Normally, the drbd
partition is supposed to accepts something lile 128 MB of data, the df
-h command showing:
Filesystem  Size        Used  Available  Use  Mounted on
/dev/drbd0   129M       4.5M  117M4%   /montagedrbd


 as given by the command
blockdev --getsize name_of_drbd_disk.

But if I upload a "too big" file (of 750 kbytes), on the partition , then
there is a failure on the node and it's written this on the screen :

node1 kernel: Assertion failure in __journal_remove_journal_head() at
journal.c:1730:"buffer_jbd(bh)"

and then in a loop (many times):

attempt to access beyond end of device
93:00 : rw=1, want=5010, limit=4976

To see where is the limit, I have put several small files (10211
bytes) on the partition to
see the limit total amount of data the partition can stand.
After the downloading files, seeing the plantage and rebooting the
machine, I can notice
that the partition has accepted until 582 Kb B. And the df command gives :
Filesystem  1k-block  Used  Available  Use  Mounted on
/dev/drbd0   131734    4697  120235     4%   /montagedrbd

The problem is I cannot see a near value among the parameters in the
following drbd.conf file :

resource drbd0 {
protocol C;

# Temps d'attente du deuxieme noeud
startup {
 wfc-timeout 10;
}

disk {
 on-io-error pass_on;
}
net     {
 timeout 60;
 connect-int 10;
 ping-int 10;
 on-disconnect reconnect;
 max-buffers 2048;
 max-epoch-size 2048;
}
syncer  {
 rate 8M; group 2;
}

on EEPCLU1 {
 device /dev/drbd0;
 disk /dev/hda7;
 address 135.9.216.42:7789;
 meta-disk internal;
}

on EEPCLU2 {
 device /dev/drbd0;
 disk /dev/hda8;
 address 135.9.216.44:7789;
 meta-disk internal;
}

}

Using the  blockdev --getsize command and indicates the size of the disk
is 136048 KB. But I cannot found in the internet how to parameter this in
the drbd.conf file. To complete the provious message, I would say when I put
a too big file on drbd partion, then before having this message in
loop on the screen:
attempt to access beyond end of device
93:00 : rw=1, want=5010, limit=4976

I had this one :
node1 kernel: Assertion failure in __journal_remove_journal_head() at
journal.c:1730:"buffer_jbd(bh)"

Also, do you know whether a particular format
(ext3|reiserfs|xfs|jfs|vfat|fat|nfs) is recommended for mounting drbd
partitions ? Now, we use the ext3 format, as specified in haresource
file (active/passive heartbeat cluster) :
EEPCLU1 135.9.216.51 \
drbddisk::drbd0 \
Filesystem::/dev/drbd0::/montagedrbd::ext3:: \
wu-ftpd

Thanks by advance !