Hp StorageWorks Scalable File Share Manuale Utente Scaricare il pdf (Pagina 41)

Starting a file system 3–9

You can expect a starting service to go to the recovering or running state after about 1 minute. The

recovering state indicates that client nodes were connected to the service before it was stopped and

those client nodes have not yet reconnected. The state of the service changes as follows, depending on the

success of the client nodes in reconnecting to the service in the period after the service starts:

• If all of the originally connected client nodes reconnect to the service successfully, the service goes to

the running state as soon as the reconnections are completed.

• If some of the original client nodes reconnect, but others fail to reconnect (because they were shut

down while the file system was stopped), the service remains in the recovering state for up to ten

minutes, and then goes to the running state.

• If none of the original clients reconnect, the service stays in the recovering state until the first new

client connection attempt. Approximately five minutes after this connection attempt, the service goes

into the running state.

When any service is in the recovering state, all client access (mount, unmount, I/O) is stalled and will

not resume until the service goes into the running state.

The timings described above are based on the default timeouts of the file system software.

3.8.1 Responding to formatting messages from the start filesystem command

Normally, the MDS and OST devices assigned to a file system are formatted when the file system is created.

However, the start filesystem command can also format MDS and OST devices when necessary.

When the start filesystem command is run, it checks whether there are file system devices that need

to be formatted; if it finds any such devices, it displays a message to allow you to decide whether to format

the devices. In the following circumstances, the command will detect devices that need to be formatted:

• If the create filesystem command failed before completing, and the devices were not all

formatted

• If additional OST devices were added to the file system (using the modify filesystem

command—see Section 5.6.1)

However, if the system database has become corrupted, the start filesystem command may report

that devices need to be formatted when in fact the devices are formatted and may have already been used

and contain data. In such cases, if MDS or OST devices are formatted unnecessarily, data will be lost. If the

start filesystem command indicates that there are unformatted devices, investigate whether there is

a valid reason why devices may need to be formatted (for example, in the circumstances listed above). If

the file system has been operating normally, and no new OST devices have been added to the file system,

you can choose to start the file system without formatting any devices.

The following is an example of the output from the start filesystem command when unformatted

devices are detected:

sfs> start filesystem scratch

Warning: the following filesystem devices do not appear to have been

formatted:

mds22

Normally this happens when a create filesystem command fails or new OSTs

have been added but not yet formatted.

If the create filesystem command failed or new OSTs are being added, you

should format the devices.

If you have already been successfully using the filesystem, and are not

adding new OSTs to the filesystem you should not format devices because

you will lose data.

Do you want to format the devices? (yes/no/cancel) [n]:

The default is no—that is, the devices will not be formatted.

1 2 ... 36 37 38 39 40 41 42 43 44 45 46 ... 361 362

Hp StorageWorks Scalable File Share Manuale Utente Pagina 41

Commenti su questo manuale

Hp StorageWorks Scalable File Share Manuale Utente Pagina 41

Commenti su questo manuale

Prodotti e manuali riguardandi Magazzini Hp StorageWorks Scalable File Share