
Creating and modifying file systems5–28
If the create filesystem command fails, you need to determine whether the file system was partially
created or not. Enter the show filesystem command; if the file system is not shown at all, this indicates
that no component of the file system has been created (that is, the create filesystem command failed
entirely); if this happens, you must correct the original problem that caused the command to fail, and then
use the create filesystem command again.
If the show filesystem command shows data about the file system but the state of the file system is not
shown as started, and/or the services are not running, the file system has been partially created.
In the following example, the show filesystem command shows that the file system has been partially
created. However, the ost36 and mds9 services are not running, and the state of the data file system is
not-started (instead of started):
sfs> show filesystem
Name State Services
------------- -------------- ----------------------------------
data not-running mds9: , ost[33-35]: running, ost36:
In such cases, you must correct the problem that caused the failure of the create filesystem command,
and then use the start filesystem filesystem_name command) to restart the file system creation
operation where it failed, as shown in the following example:
sfs> start filesystem data
While the start filesystem command is working, closely observe the messages printed by the
command. In addition, examine event log messages from the lustre facility, as explained above.
5.5.1 If the create filesystem command times out while configuring file system
services
The create filesystem command sometimes times out while the file system services are being
initialized.
• If the command times out while the MDS service is being initialized, output similar to the following is
shown:
Failed to configure MDS mds4, command timed out
• If the command times out while an OST service is being initialized, output similar to the following is
shown:
Failed to configure OSTs on south3, command timed out
When this happens, the show filesystem command shows the file system as partially created—that is,
the file system exists, but the state of the file system is shown as not-running.
To recover from this problem, increase the value of the lustre.mkfs_timeout attribute in the
hpls_attributes table. Change the value from the default of 3600 seconds to 7200 seconds, as
follows:
mysql> update hpls_attributes set val=’7200’ where attribute=’lustre.mkfs_timeout’;
When you have changed the lustre.mkfs_timeout attribute, enter the start filesystem
filesystem_name command again, as shown in the following example, where the file system is called
data:
sfs> start filesystem data
NOTE: If the create filesystem command fails for reasons other than that the command timed out
while initializing the file system services, output similar to the following is shown:
Failed while waiting to configure the MDS/ADM for data_adm
Or:
Failed while waiting to configure OSTs on south3
In such cases, changing the lustre.mkfs_timeout attribute will not solve the problem, and you must
look for the root cause of the failure of the create filesystem command,as described in Section 5.5.
Comentários a estes Manuais