Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
Jerry Amundson wrote:
> I would think your storage could handle the 110M rate, but have you checked?
> http://lists.linbit.com/pipermail/drbd-user/2008-July/009897.html
I have not run any benchmarks (yet), however this was only done to
sync the nodes up, and was up to date at the time of the crash.
Syncs running at 100M complete fine.
To reiterate, I was performing an rsync from a remote host on the
WAN to the primary node at our data center. Given the much slower
rate of transfer across the WAN, I can't imagine the node sync rate
being a factor here.
> Maybe the drbd.conf would be useful...
global {
usage-count yes;
}
common {
protocol C;
}
resource data {
on haws1 {
device /dev/drbd0;
disk /dev/mapper/VolGroup00-LogVol02;
address 10.0.0.222:7789;
meta-disk internal;
}
on haws2 {
device /dev/drbd0;
disk /dev/mapper/VolGroup00-LogVol02;
address 10.0.0.223:7789;
meta-disk internal;
}
}
> I doubt 2.6.18_92.1.10 would make a difference, but it should be there now.
> Why centosplus - because of 3ware?
Because I prefer to install everything as an RPM, even if I need to
build them myself. centosplus already provides the kernel module and
software rpms (no need to reinvent the wheel), however it's for
their kernel which includes some things I really don't need. That
said I could use their SRPM to build the module against the stock
kernel, but before I do that, I'd really like to understand the
issue reported in the kernel panic.
I can't believe I'm the first person to run into this, but if so,
then there's something specific about my configuration that must be
the cause. That said, what could it be? In my configuration I'm using :
CentOSPlus kernel, DRBD module, and DRBD build
DRBD on an LVM Partition
3Ware 9650 SATA 2 controller
750GB disks
3w_9xxx driver provided by CentOSPlus kernel
Quad Core Xeon 3k series (SMP x 4 cores)
4GB of memory
So the possibilities could be the DRBD build itself, the kernel, the
disc controller driver, the disk size (addressing issue?), the
partition layer (LVM), a hardware issue, or a bug in DRBD.
The systems are brand new and were put through an extensive 48 hour
burn-in process by the hardware vendor, that consists of repetitive
hardware tests. I really don't think this is an issue, but I won't
rule it out.
Unfortunately I don't have an third machine with this configuration
to install differently. If it came down to it I'd consider
re-installing the second node, but I really don't have the time to
take a lot of shots in the dark. I will for now upgrade the kernel
and DRBD module and attempt another transfer.
Regards,
Chris