Recent content by David Calvache Casas

Hi guy After destroying the osds, it doesn't allow me to rejoin them because it tells me that it is a member of an LVM group. but neither lvscan, nor vgscan nor ceph lvm ls shows said membership. The OSD Add buton doesn´t display the disk to add. This show the membership. but, this tells...

I have realized that my DB partitions are sized to 1GB (my bad). I have one SSD 480 Gb for two DB/wall partitions that belongs to two 6TB hard drives. After a warning of spillover, i discovered the 1 gigabyte partitions. Then my plan was resize to 220Gb each, but because the spillover that is...

10.0.1.1 is the default gateway for your network?. Then this is the gateway for your vms, if they are in that 10.0.1.0/24 net

Samsung SM883, they perform well.

Yes, I think so. I remember somebody going to downsizing the replication numbers. But doing that ypu put yourself in risk.

2/1 in a test enviroment it's OK, In a production system it's a noway. If you lost one one, nothing happens, but if something occurs before the rebuild, you are game over. Be ready for a mass restauration for backups, because you will have data losses for sure. With 3/2 , if you lost a second...

please post : ceph -s ,ceph health detail , pvecm status. So yo dont have to rebuild the pool, just increment the replica size to 3. Something like ceph osd pool set POOL_NAME size 3

Gotcha!!! It was a parameter : mon osd min in ratio = 0.75, modifing this to 0.70 permit me to lost more OSD and mark it as Out. Thank you for your help.

Can anybody test this? I'm testing this: 1.- If one or several OSD in one HOST goes down in a host, they work as expected, marked down, and afeter 600seg marked as out. The ceph rebuilds itself. 2.- If one host goes down, the all OSD in that host are marked as down, and after 600seg marked...

Yep, I know. But one of the OSD never goes to down/out (from down/in). All the other became marked as out, and start recovery, but one osd still faulty.

Hi. I have a problem. I am testing a 7 node cluster. Each node has 1 nvme (4 OSD) and two hd (2 OSD), so 6 OSD each node. There are two replication rules (type nvme and hd) and two pools (fast and slow) acording to the rules. All its ok, but when i shutdown one one, and thereafter another...

OK, Issue resolved. Was my fault Igmp querier not enabled in vlan at the switch. I changed the vlans and forget to enable it. I apologize for the inconvenients caused. Thank you very much for the waste of time and brain.

The 10 minute test is Ok too. 10.9.5.151 : unicast, xmt/rcv/%loss = 600000/600000/0%, min/avg/max/std-dev = 0.023/0.090/0.918/0.032 10.9.5.151 : multicast, xmt/rcv/%loss = 600000/600000/0%, min/avg/max/std-dev = 0.023/0.093/0.920/0.032 10.9.5.152 : unicast, xmt/rcv/%loss =...

I will test the omping for 10 minutes... there is no bridge in corosync interfaces.... more /etc/network/interfaces auto lo iface lo inet loopback auto enp175s0f0 iface enp175s0f0 inet static address 10.9.5.156 netmask 255.255.255.0 #Corosync RING0 auto enp175s0f1 iface...