Thank you for your reply
I hava check both vm host and physical host, Free Ram is enough (100G left) And load 0.5 （5min 32cores）
in short . Cpu and memory are sufficient. Iops （see ceph -s） only 2000op/s. (16 ssd disk cache , 92 hdd disk)
The physical host is centos7 . Kernel version 3.10.0. openstack version:pike . Ceph version luminous . Bluestore
I hava 5 pools . Volumes ,images, volume_cache, vms, vms_cache. volume_cache is ssd cache tier of Volumes and vms_cache is ssd cache tier of vms.
I notice that you said ceph cache. Do you mean ceph rbd cache ？
Did you check your KVM host RAM usage?
We saw this on host very very loaded with overcommit in RAM causes a random crash of VM.
As you said for solve must be remounted externaly and fsck. You can prevent it disabled ceph cache at Openstack Nova host. But your VM’s are going get less performance.
Whats you Ceph & Openstack version?
Regards
De: ceph-users <ceph-users-bounces@xxxxxxxxxxxxxx>
En nombre de ??
Enviado el: viernes, 17 de mayo de 2019 9:01
Para: ceph-users <ceph-users@xxxxxxxxxxxxxx>
Asunto: [ceph-users] openstack with ceph rbd vms IO/erros
hi:
I hava a openstack cluster with a ceph cluster ，use rbd，ceph cluster use ssd pool tier.
some vm on openstack sometimes crashed in two case .
1. become readonly filesystem. after reboot ,it work fine again.
2. IO errors . I must repair the file system by fsck. thenreboot , it work fine again.
I do not know if this is ceph bugs or kvm bugs.
I need some ideas to resolv this , Anyone can help me ?
Look forward to your reply
_______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
