Shulou(Shulou.com)06/02 Report--

These commands are uploaded to the CephFS file system in your Ceph cluster. Note that only one file system is allowed by default. To enable the creation of multiple file systems, use the ceph fs flag set enable_multiple true command

File system related

Ceph fs new

This command creates a new file system, and the file system name and metadata pool name are clear. The specified datapool is the default datapool, which cannot be changed once set, and each file system has its own set of MDS daemons assigned to the level, so make sure you have enough backup daemons to accommodate the new file system.

Ceh fs ls

List existing file systems

Ceph fs dump [epoch]

This will dump the FSMap for a given period of time (default: current), including all file system settings, the MDS daemon and its levels, and a list of standby MDS daemons

Ceph fs rm [--yes-i-really-mean-it]

"destroy the CephFS file system, which erases information about the file system status from FSMap, and the metadata pool and datapool remain the same and must be destroyed separately."

Ceph fs get

Get information about named file systems, including settings and levels, which is a subset of the same information from the fs dump command

Ceph fs set

Change settings on the file system that are specific to named file systems and do not affect other file systems

Ceph fs add_data_pool

Add a datapool to the file system, which can be used for file layouts as an alternate location to store file data.

Ceph fs rm_data_pool

This command removes the specified pool from the list of data pools in the file system, and if any file has the layout of the deleted data pool, the file data is not available and the default data pool (the first data pool created when the file system is created) cannot be deleted.

Set correlation

Ceph fs set max_file_size

CephFS has a configurable maximum file size, which is 1TB by default, and if you want to store large files in CephFS, you can set this limit even higher, which is a 64-bit field. Setting max_file_size to 0 does not disable this restriction, it only restricts the permission of the client to create empty files.

Maximum file size and performanc

CephFS enforces the maximum file size limit when appending files or setting file sizes, which does not affect anything stored in any storage mode. When a user creates a large file (without having to write any data to it), some actions (such as deletion) cause MDS to perform a large number of operations to check for the existence of possible RADOS objects (depending on the file size)

The max_file_size setting prevents users from creating exabyte-sized files, causing load on the MDS because it attempts to enumerate objects during operations such as counting or deleting.

Down dropped the cluster

Shut down the CephFS cluster by setting the down parameter

Ceph fs set down true

Restore cluster

Ceph fs set down false

This also restores the previous values of max_mds, shuts down the MDS daemon in some way to flush journals to the metadata pool, and stops I / O operations on all clients.

Quickly shut down the cluster for deletion or disaster recovery

To allow quick deletion of a file system (for testing) or to quickly close a file system and MDS daemon, use the fs fail command, which sets the file system flag to prevent alternate files from being activated on the file system

Ceph fs fail

You can also do this manually by doing the following

Ceph fs set joinable false

The operator may then fail all ranks, which will cause the MDS daemon to be reborn as a standby and the file system to be degraded

Ceph mds fail:

Once all ranks is inactive, you can also delete the file system or leave it in that state for other purposes (perhaps disaster recovery). To restore the cluster, simply set the joinable flag

Ceph fs set joinable true

Daemon process

Most commands that operate on mds take parameters, which can take one of three forms

:

:

Commands to operate the MDS daemon

Ceph mds fail

Mark the MDS daemon as lost, which means that the cluster will perform this operation when the MDS daemon fails to send a MDS_beacon_grace second message to mon. If the daemon is active and a suitable standby server is available, use mds fail to force a failover to the standby server.

If the MDS daemon is actually still running, using MDS fail will cause the daemon to restart. If it is active and a standby is available, the "failed" daemon will be returned as a standby

Ceph tell mds. Command...

Send the command to the MDS daemon, using mds. Send commands to all daemons, using ceph tell mds. To view the help documentation

Ceph mds repaired

Get metadata for a given MDS that is known about Mon

Ceph mds repaired

Mark the file system level as repaired. As the name implies, this command does not change MDS, it operates on file system ranks that has been marked as corrupted

Minimum client version

Sometimes it is necessary to set the minimum Ceph version of the client to function properly in order to connect to the CephFS cluster. Older clients may sometimes still be running with errors that can cause locking problems between clients (due to feature releases). CephFS provides a mechanism for setting the minimum client version

Ceph fs set min_compat_client

For example, only Nautilus clients are allowed to use

Cephfs set cephfs min_compat_client nautilus

Clients running older versions will be automatically expelled

Global Settin

Ceph fs flag set []

Set the global CephFS flag (that is, not specific to a specific file system). Currently, the only flag setting is "enable_multiple", which allows multiple CephFS file systems. Some flags require you to use "- yes-i-really-mean-it" or a similar string to prompt for confirmation. Please consider these operations carefully before continuing, as they are more dangerous operations.

Advanced usage

These commands are not required in normal operation and can be used in special cases. Incorrect use of these commands can cause serious problems, such as file system inaccessibility

Ceph mds compat rm_compat # remove compatibility feature Flag

Ceph mds compat rm_incompat # removes incompatible feature flags

Ceph mds compat show # displays the MDS compatibility flag

Ceph mds rmfailed # remove rank from the failed collection

Ceph fs reset # this command resets the file system state to the default name (except name and pool), and non-zero rank is saved in the stop set

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.