cephmap相关crushmap操作cephosdgetcrushmap-ocrushmap_compiled_filecrushtool-dcrushmap_compiled_file-ocrushmap_decompiled_filecatcrushmap_decompiled_file*好像crushmap和osdmap不完全一样root@node1:~#cephosdgetcrushmap-ocrushmap_compiled_filegotcrushmapfromosdmapepoch73root@node1:~#cephosdgetmap-oosdmapgotosdmapepoch73monmap操作rbd相关的ceph配置文件调整##从6.0升级到之后需要关于mon的配置6.1修改，否则可能导致ceph-s输出卡住。monpgwarnmaxperosd=10240monsyncmaxpayloadsize=4096monpgwarnmaxobjectskew=0cephdebugdebug设置后需要记得关闭，否则日志会太大。开启osd调试，运行时调试，调整配置文件调试设置，保证重启后仍然有效配置文件调整中有一个格式debugmon=0/10debugpaxos=0/10格式说明：内存日志级别/fileloglevel在osd启动前运行脚本并通过启动前命令参数指定脚本[osd.8]host=10.46.200.13publicaddr=10.46.200.13clusteraddr=10.46.200.13filestoremaxsyncinterval=3osdjournal=/dev/disk/by-partlabel/85-osd-3-journalosduuid=08119283-4b03-469d-a72f-5877commandstart=pre/usr/local/bin/loop_osd_pre_start.sh8||true##加true表示如果脚本执行失败，会继续启动ceph运行时调参。Runtime参数调整和配置文件参数调整，参数的写法不一样，Runtime需要下划线。运行时写法如下"osd_op_thread_timeout":"15","osd_recovery_thread_timeout":"30","osd_command_thread_timeout":"600","filestore_op_thread_timeout":"60","filestore_op_thread_suicide_timeout":"180",ceph.conf的写法方法如下：osdopthread_timeout=150osdrecoverythreadtimeout=300osdcommandthreadtimeout=6000filestoreopthreadtimeout=600filestoreopthreadsuicidetimeout=1800cephosdsetcephosdsetnoout#集群停止恢复，503cephosdsetnorecover504cephosdsetnobackfillcephosdunsetnooutceph参数记录控制jounal写入底层数据的时间间隔有时，性能底层磁盘不足root@Storage-03:/var/log/ceph#cephdaemonosd.8configshow|grepinterval|grepsync"filestore_max_sync_interval":"10",
