it depends heavily on the Windows Version, any version before Server 2012 or 8 does not support the scsi Unmap command
Also it does not clean up immediately, you should also look for this:
fsutil behavior query disabledeletenotify
If the parameter is set to "1" trim/unmap is disabled and...
if you move with P2V the empty blocks are copied over too.
You should probably set "discard" in the disk definition, so you can regain the zero blocks.
But be aware: this works only with Windows 10 or 2012 up
Windows 7/2008R2 only support discard with ATA not with SCSI
if you want to extend a RaidZ2, you have to extend it by a same size VDEV.
So with 4 * 1TByte RaidZ2 you have to add again a VDEV with 4 * TByte.
But: Z2 means 2 drives for redundancy. From a capacity point of view it is not better than 2 Mirror VDEV's
Z2 makes sense for high capacity with at...
it is supported very fine. the management is simple. I run 4 Clusters each with CEPH
As you are versed in CEPH you should know what to do to get it running fast. Just keep off the fingers from rotating disks and use a fast backend network >= 10 GBit/sec
I do not understand why you want a separate Journal for a full flash deployment. if your journal SSD fails you loose immediately all 6 OSD's depending on it. Also the write load on the journal will be 6 times the OSD's
compare: 0.8 DWPD to 3 DWDP -> so if you are really hitting the DWDP Limit...
Sorry, just hit the post button to early, now here the relevant information:
Write Configuration When Not Quorate
If you need to change /etc/pve/corosync.conf on an node with no quorum, and you know what you do, use:
pvecm expected 1
This sets the expected vote count to 1 and makes the cluster...
Very curious, we have many disks of different types and vendors running with ZFS since years (RaidZ2 and Mirror). Of course we see disks passing away, but with normally expected rates ( usually < 1 per month out of around 500 disks ). So I bet your hoster has just no luck with these disks.
It depends heavily on the switch you use. With an dumb switch there is just Active/passive possible (anything else will not work reliable).
you need a managed switch on which you can setup LACG groups (or you can setup a full mesh with direct connections, openvswitch can do that) _but_ don't...