TOPIC:

Wrong forum.

Did you build the config by hand or use the installer?
If you built by hand, double check the LVM filters in lvm.conf AND ..master/lvm.conf.

Also, if you have not done so already, you should erase any cached LV data.

Both of these would cause the condition you are seeing.

If you built using the automated installer then there is something else wrong. Let me know how it was built so we know what direction to take this.

I actually copied the lvm.conf and master/lvm.conf from a working deployment ; device names are the same. And yes, I deleted the cache and rescanned.
I'll definitely double check, though.

Confirmed. In both /etc/lvm/lvm.conf and /etc/lvm/master/lvm.conf, I have the following at line 142:
filter = [ "r|/dev/xvd.|", "r|/dev/VG_Xen.*/*|", "r|/dev/sdb|", "r|/dev/drbd.*|"]


/dev/sdb is the physical disk that drbd is accessing, with /dev/drbd1 being the device.

There is no /etc/lvm/cache/.cache file ; after deleting it originally and running vgscan, it was never re-created. I verified this on another known working pool.

Sorry I missed it earlier ; I built by hand.

if you are comfortable that your configs are correct, then it sounds like the primary node may be read only for some reason. You can try stopping iscsi-ha, its watchdog, tgtd, DRBD. Then start iscsi-ha again to ensure that it brings up the target and DRBD in the correct order.

Is there a method to check that's the case before stopping services? I've rebooted both nodes several times and still have the same issue...
I'm reasonably sure my configs are correct ; I've double checked them multiple times, and it's not my first build.