man page(1) manual page

options is used to change configurable filer software options. If no options are specified, then options prints the current value of all available options. The default value for most options is off, which means that the option is not set. Changing the value to on enables the option; for most options, the only valid values are on (which can also be expressed as yes, true, or 1) in any mixture of upper and lower case, and off (which can also be expressed as no, false, or 0) in any mixture of upper and lower case. The description of the option will indicate the default if it is not off, and will indicate what values are allowed if it isn't an on/off option. If it is desired to make an option setting permanent, the necessary options command must be placed in the /etc/rc file, as options settings are not preserved across system reboots. The legal options are as follows:

autosupport.enable
Enables/disables the autosupport email notification feature (see na_autosupport(8)). The default is on to cause autosupport mail to be sent.

autosupport.mailhost
Defines the list of up to 5 mailhost names. The host names should be entered as a comma separated list with no spaces in between. The default is an empty list.

autosupport.to Defines the list of recipients for the autosupport email notification. Up to 5 mail addresses are allowed. Enter the addresses as a comma separated list with no spaces in between. The default is autosupport@netapp.com.

autosupport.noteto
Defines the list of recipients for the autosupport short note email. Up to 5 mail addresses are allowed. Enter the addresses as a comma separated list with no spaces in between. The default is a null list to disable short note emails.

autosupport.from

Defines the user to be designated as the sender of the notification. The default is autosupport.

autosupport.doit
Triggers the autosupport daemon to send email notification immediately. A text word entered as the option is sent in the notification subject line and should be used to explain the reason for the notification.

cifs.generic_account
The cifs.generic_account option can enable a user who has no password account in /etc/passwd on the filer or NIS to get access to the filer, provided that the filer uses a Domain Controller for authentication and the user is in a trusted domain. If the option is set to the name of an account, the user gets the UNIX user ID, group ID, and group set of that account. If the option is blank, generic access is disabled.

cifs.guest_account
The cifs.guest_account option can enable a user to get access to the filer provided that either the filer uses a Domain Controller for authentication and the user is not in a trusted domain, or the filer uses /etc/ or NIS for authentication and the user has no entry in /etc or NIS. If this option is set to the name of an account in the password database, a user logging into the filer will be assigned to the guest account if their name is not listed in the password database (when using /etc or NIS) or if the user is not from a trusted domain (when using a domain controller). The configured user name will be used for the UNIX user ID, group ID, and group set of the specified account. If the option is blank, guest access is disabled.

cifs.home_dir When set to the pathname of a directory, this defines the path to the ``homes directory''. The directories under this path should have the names of users as their names. When a user connects to the filer using CIFS and there is a directory name that exactly matches the user name, they will see a share of that name (truncated to 12 characters) that is their ``home directory". Only the user can access the home directory using this share. All other users are denied access.

cifs.oplocks.enable
When cifs.oplocks.enable is on (the default), the filer allows clients to use oplocks (opportunistic locks) on files. Oplocks are a significant performance enhancement, but have the potential to cause lost cached data on some networks with impaired reliability or latency, particularly wide-area networks. In general, this option should be disabled only to isolate problems.

cifs.scopeid NetBIOS scope IDs allow the system administrator to create small workgroups out of a network by partitioning the NetBIOS name space; only clients with the same NetBIOS scope ID as the filer will be able to use the filer as a CIFS server. Normally, the scope ID is a null string, but if the filer is to run in a NetBIOS scope other than the default one, its scope ID must be set to the scope ID of that scope. The scope ID can be changed only when CIFS is not running.

cifs.symlinks.cycleguard
The cifs.symlinks.cycleguard option (on by default), eliminates the possibility of traversing directories cyclically during the following of symbolic links. With this option set to on, the filer does not follow symbolic links that contain the "dot" (`.') or "dot-dot" (`..') component -- symbolic links that could refer to a directory higher in the same tree. With this option set to off many standard windows apps (such as Find in Win95 / NT4.0) will not operate correctly when a symlink points to a parent directory. This is because they do not understand symbolic links and will repeatedly loop on them. Users should use caution when changing this option.

cifs.symlinks.enable
When cifs.symlinks.enable is on (the default), if the object being accessed by a CIFS client is a symbolic link (whether absolute or relative), the filer follows the link with the proviso that the ultimate target turns out to reside within the originating share (thus ensuring that the client has access permission to the target).

df_2gb_lim: Causes the filer to return replies to the "file system statistics" NFS request that show no more than (2**31)-1 (or 2,147,483,647) total, free, or available bytes (i.e., 2GB) on the file system.

Some NFS clients require this option because, if they get return values from the "file system statistics" request with more than the specified number of bytes, they'll incorrectly compute the amount of free space on the file system, and may think that there's no free space on a file system that has more than 2GB free.

dns.domainname Sets the DNS domainname to the specified domainname.

dns.enable: Enables DNS client on the filer. The DNS domain must be set and the /etc/resolv.conf file must exist prior to enabling DNS.

ip.path_mtu_discovery.enable
Enables/disables path MTU discovery; it is currently used only by TCP. Path MTU discovery, described in RFC 1191, allows a host to discover the ``maximum transmission unit'', i.e. the largest link-level packet that can be transmitted, over a path from that host to another host. This means that the filer needn't choose a conservative packet size for a TCP connection to a host not on the same net as the filer, but can attempt to discover the largest packet size that can make it to the other host without fragmentation.

httpd.enable: Enables HTTP access to the filer.

httpd.admin.enable
Enables HTTP access to the administration area of the filer, via a private Network Appliance URL: any URL beginning with /na_admin is mapped to the directory /etc/http. Thus, a man page on the filer toaster with the file name /etc/http/man/name can be accessed with the URL http://toaster/na_admin/man/name.

httpd.log.max_file_size

Specifies the maximum size that the HTTP log file /etc/log/httpd.log can grow to. The default is 2147483647, which is the largest file size that many clients support.

httpd.rootdir Specifies the complete pathname of the root directory that contains files and subdirectories for HTTP access.

httpd.timeout Specifies the minimum amount of time (in seconds) before an idle HTTP connection will time out. The default is 900 seconds, which is fifteen minutes.

httpd.timewait.enable
When enabled, the filer will put HTTP connections which have been closed by the client into the TIME_WAIT state for one minute, which is twice the maximum segment lifetime (2*MSL). By default, TIME_WAIT state is bypassed for HTTP connections.

minra: Configures the filer to perform minimal read ahead. By default, the filer does very aggressive read ahead.

mount_rootonly When enabled, the mount server will deny the request if the client is not root user using privileged ports. By default, the feature is enabled for more secure access.

nfs.per_client_stats.enable
Enables/disables the collection and display of per-client NFS statistics as described in na_nfsstat(1).

nfs.tcp.enable When enabled, the NFS server supports NFS over TCP. By default, the feature is enabled; it can be disabled if there is a problem with some client when using NFS over TCP, and that client cannot be configured to use NFS over UDP.

nfs.v3.enable When enabled, the NFS server supports NFS version 3. By default, the feature is enabled; it can be disabled if there is a problem with some client when using NFS version 3, and that client cannot be configured to use NFS version 2.

nfs.webnfs.enable
When enabled, the NFS server supports WebNFS lookups. By default, WebNFS lookups are disabled.

nfs.webnfs.rootdir
Specifies the WebNFS rootdir. Once the rootdir is set, WebNFS clients can issue lookups relative to the rootdir using the public filehandle.

nfs.webnfs.rootdir.set
After specifying the rootdir, this option needs to be enabled for the rootdir setting to take effect. Disabling this option disables the existing rootdir setting.

nis.domainname Sets the NIS domain to the specified domainname.

nis.enable: Enables NIS client on the filer. The NIS domain must be set prior to enabling NIS.

no_atime_update
This prevents the update of the access time on an inode when a file is read. This option is useful for servers with extremely high read traffic, since it prevents writes to the inode file from contending with reads from other files. It should be used carefully, i.e., when it is known in advance that the correct access time for inodes will not be needed.

nosnap: Temporarily disables automatic snapshots.

nosnapdir: Disable the visible .snapshot directory that is normally present at client mount points, and turn off access to all other .snapshot directories.

nfs.big_endianize_fileid
Put the file ID in the file handle reported by the server in ``big-endian'' form. This helps improve performance on directory lookup operations on certain clients that use the file ID within the file handle as a hash key.

pcnfsd.enable Enables/disables the PCNFSD (PC)NFS authentication request server (see na_pcnfsd(8)).

raidtimeout: Set the time in hours, as a number greater than or equal to 1, that the system will run after a single disk failure has caused the system to go into degraded mode. The default is 24. If the raidtimeout option is specified after the system is already in degraded mode, the timeout is set to the value specified and the timeout restarted.

raid.reconstruct_speed
Specifies the speed at which the RAID reconstruction should occur ranging from the slowest speed 1 to the fastest speed possible 10. The RAID reconstruction process is given more cpu time as the speed is increased, so increasing the speed of the reconstruction will take away cpu time for network operations. The default speed is 4 which is roughly 40% of the cpu time, though more time may be used if there is idle time available.

raid.scrub.enable
Enables/disables the RAID scrub feature (see na_raid(1)). By default, it is enabled. This option only affects the scrubbing process that gets started from cron. For user requested scrubs, this option is ignored.

root_only_chown
When enabled, only the root user can change the owner of a file. When disabled, non-root users can change the owner of files that they own. When a non-root user changes the owner of a file they own, both the set-UID and set-GID bits of that file are cleared for security reasons. A non-root user is not allowed to give away a file if it would make the recipient overrun its user quota. root_only_chown is enabled by default.

telnet.hosts Allows specifying up to 5 clients that will be allowed telnet access to the server. The host names should be entered as a comma separated list with no spaces in between. Enter a "*" to allow access to all clients; this is the default. Enter a "-" to disable telnet access to the server.

wafl.maxdirsize
Sets the maximum size (in K-Bytes) that a directory can grow to. This is set to 10240 by default; it limits directory size to 10MBytes and can hold over 300,000 files. Most users should not need to change this setting. This option is useful for environments where system users may grow a directory to a size that starts impacting system performance. When a user tries to create a file in a directory that is at the limit the system returns a ENOSPC error and fails the create.

Multiple options can be set at once in an options command. For example:

toaster> options minra on nosnap on nosnapdir on raidtimeout 48

sets minra to on, sets nosnap to on, sets nosnapdir to on, and sets raidtimeout to 48.

BUGS

A perfect appliance would need no options (other than, perhaps, a darkness adjustment knob). Though, user Nigel Tuffnell reports he likes the raid.reconstruct_speed knob that goes from 1 to 10, but he requests a future enhancement to "go to 11 because it's one better, isn't it?"

Table of Contents

NAME
SYNOPSIS
DESCRIPTION
SEE ALSO
BUGS

NAME

SYNOPSIS

DESCRIPTION

SEE ALSO

BUGS