Well. it seems memory
I have 3 ODS per host with 8G RAM and block.db in SSD
Setting bluestore_cache_size_ssd=1G seems to have fixed the problem. No
new inconsistencies.
On 21/08/18 16:09, Paul Emmerich wrote:
Are you running tight on memory?
Paul
2018-08-21 20:37 GMT+02:00 Alfredo Daniel Rezinovsky
<alfredo.rezinovsky@xxxxxxxxxxxxxxxxxxxxxxxx>:
My cluster suddenly shows many inconsistent PGs.
with this kind of log
2018-08-21 15:29:39.065613 osd.2 osd.2 10.64.1.1:6801/1310438 146 : cluster
[ERR] 2.61 shard 5: soid 2:864a5b37:::1000070510e.00000004:head candidate
had a read error
2018-08-21 15:31:38.542447 osd.2 osd.2 10.64.1.1:6801/1310438 147 : cluster
[ERR] 2.61 shard 5: soid 2:86783f28:::10000241f7f.00000000:head candidate
had a read error
Al error fixes with "ceph pg repair" eventually but new inconsistencies
appears.
smart and kernel logs shows no hdd problems.
I have bluestore OSDs in HDD with journal in an SDD partition.
--
Alfredo Daniel Rezinovsky
Director de Tecnologías de Información y Comunicaciones
Facultad de Ingeniería - Universidad Nacional de Cuyo
_______________________________________________
ceph-users mailing list
ceph-users@xxxxxxxxxxxxxx
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
--
Alfredo Daniel Rezinovsky
Director de Tecnologías de Información y Comunicaciones
Facultad de Ingeniería - Universidad Nacional de Cuyo
_______________________________________________
ceph-users mailing list
ceph-users@xxxxxxxxxxxxxx
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
