Introduction

This appendix contains the error and log messages displayed by the metadevice utilities of DiskSuite 4.0.

Errors that deal with command usage and other simple error messages are not documented in this appendix. All DiskSuite error messages are displayed in the following format:

program name: host: [optional1:] name: [optional2]: error message...

where:

• program name: is the name of the application name and version being used (for example, DiskSuite 4.0).
• host: is the host name of the machine on which the error occured (for example, blue).
• [optional1]: is an optional field containing contextual information for the specific error displayed (ie. mountpoint or which daemon returned the error).
• name: is the command name which generated the error message (for example, metainit).
• [optional2]: is a second optional field containing additional contextual information for the specific error displayed
• error message... is the error message itself (as listed in this appendix).

For the purpose of this appendix, only the final portion (error message...) of each error message is listed.

The log messages listed near the back of this appendix are divided into three categories:

• Notice log messages
• Warning log messages
• Panic log messages

Error Messages

The error messages displayed by DiskSuite are listed in alphabetical order below. The message is preceded by some or all of the variables described in the previous section. Other variables included in these messages indicate the following:

• nodename is the name of a specific host.
• drivename is the name of a specific drive.
• metadevice is the number of a specific metadevice device or hot spare pool.
• setname is the name of a specific diskset.
• num is a number.

add or replace failed, hot spare is already in use

The hot spare that is being added or replaced is already in the hot spare pool.

administrator host nodename can't be deleted, other hosts still in set. Use -f to override

The host which owns the diskset cannot be deleted from the diskset without using the -f option to override this restriction. When the -f option is used, all knowledge of the diskset is removed from the local host. Other hosts within the diskset are unaware of this change.

administrator host nodename deletion disallowed in one host admin mode

The administrator host is the host which has executed the command. This host cannot be deleted from the diskset if one or more host in the diskset are unreachable.

already has log

The specified metatrans device already has an attached logging device.

already used in metadevice

The specified component is currently being used in the metadevice.

attempt to detach last running submirror

An attempt was made to detach the last submirror. The operation would result in an unusable mirror. DiskSuite does not allow a metadetach to be performed on the last submirror.

attempt an operation on a submirror that has erred components

An attempt was made to take a submirror offline or detach a submirror that contains the data. The other submirrors have erred components. If this operation were allowed, the mirror would be unusable.

attempt an operation on a submirror in illegal state

An attempt was made to take a submirror offline that is not in the OKAY state or to online a submirror that is not in the offlined state. Use the -f option if you really need to offline a submirror that is in a state other than OKAY.

attempt to replace a component on the last running submirror

An attempt was made to replace a component in a one-way mirror.

attempted to clear mirror with submirror(s) in invalid state

The user attempted to use the metaclear command on a metamirror that contained submirrors that weren't in the OKAY state (Needs maintenance state). If the metamirror must be cleared, the submirrors must also be cleared. Use -r (recursive) to clear all the submirrors, or use -f (force) to clear a metamirror containing submirrors in the Needs maintenance state.

can't attach labeled submirror to unlabeled mirror

An attempt was made to attach a labeled submirror to an unlabeled mirror. A labeled metadevice is a device whose first component starts at cylinder 0. To prevent the submirror's label from being corrupted, DiskSuite does not allow labeled submirrors to be attached to unlabeled mirrors.

can't find component in unit

An attempt was made to replace or enable a component that did not exist in the specified metadevice.

can't find submirror in mirror

An attempt was made to either metaonline(1M), metaoffline(1M), or metadetach(1M) the submirror, dnum. The submirror is not currently attached to the specified metamirror causing the command to fail.

can't include device dev, it already exists in dnum

An attempt was made to use the device dev in a new metadevice and it already is used in the metadevice dnum.

can't include device dev, it overlaps with a device in dnum

The user has attempted to use device dev in a new metadevice which overlaps an underlying device in the metadevice, dnum.

cannot delete the last database replica in the diskset

An attempt was made to delete the last database replica in a diskset. To remove all database replicas from a diskset, delete all drives from the diskset.

cannot enable hotspared device

An attempt was made to perform a metareplace -e (enable) on an underlying device which is currently hot spared. Try enabling the hot spare component instead.

can't modify hot spare pool, hot spare in use

An attempt was made to modify the associated hot spare pool of a submirror, but the submirror is currently using a hot spare contained within the pool.

checksum error in mddb.cf file

The /etc/opt/SUNWmd/mddb.cf file has probably been corrupted or user-edited. The checksum this file contains is currently invalid. To remedy this situation: 1. Delete the mddb.cf file. 2. Delete a database replica. 3. Add back the database replica.

component in invalid state to replace \ - replace "Maintenance" components first

An attempt was made to replace a component that contains the only copy of the data. The other submirrors have erred components. If this operation were allowed, the mirror would be unusable.

data not returned correctly from disk

After a replica of the state database is first created, it is read to make sure it was created correctly. If the data read does not equal the data written this message is returned. This results from unreported device errors.

device not in set

An attempt was made to use a component for a shared metadevice or shared hot spare pool whose drive is not contained within the diskset.

device in shared set

An attempt was made to use a component for a local metadevice or local hot spare pool whose drive is contained within the diskset. The drives in the local diskset are all those which are not in any shared disksets.

device is too small

A component (dev) in stripe num is smaller than the interlace size specified with the -i flag in the md.tab file.

device size num is too small for metadevice database replica

An attempt was made to put a database replica on a partition that is not large enough to contain it.

devices were not RAIDed previously or are specified in the wrong order

An attempt was made to metainit a RAID device using the -k option. Either some of the devices were not a part of this RAID device, or the devices were specified in a different order than they were originally specified.

drive drivename is in set setname

An attempt was made to add the drive drivename to a diskset which is already contained in the diskset setname.

drive drivename is in use

An attempt was made to add the drive drivename to a diskset, however a slice on the drive is in use.

drive drivename is not common with host nodename

An attempt was made to add the drive drivename to a diskset, however, the device name or device number is not identical on the local host and the specified nodename; or the drive is not physically connected to both hosts.

drive drivename is not in set

An attempt was made to delete the drive drivename from a diskset and the diskset does contain the specified drive.

drive drivename is specified more than once

The same drive (drivename) was specifed more than once in the command line.

driver version mismatch

The utilities and the drivers are from different versions of the DiskSuite package. It is possible that either the last SUNWmd package added did not get fully installed (try running pkgchk(1M)), or the system on which DiskSuite was recently installed has not been rebooted since the installation.

failed to take ownership of a majority of the drives

Reservation of a majority of the drives was unsuccessful. It is possible that more than one host was concurrently attempting to take ownership of the same diskset. One host will succeed, and the other will receive this message.

growing of metadevice delayed

The attempted growth of a submirror has been delayed until a mirror resync finishes. The metamirror will be grown automatically upon completion of the resync operation.

has a metadevice database replica

An attempt was made to use a component (ie. for a hot spare) which contains a database replica.

host nodename already has a set numbered setnumber

An attempt was made to add a host nodename to a diskset which has a conflicting setnumber. Either create a new diskset with both hosts in the diskset, or delete one of the conflicting disksets.

host nodename already has set

An attempt was made to add a host nodename to a diskset which has a different diskset using the same name. Delete one of the disksets and recreate the diskset using a different name.

host nodename does not have set

An attempt was made to delete a host or drive from a set, but the host nodename has an inconsistent view of the diskset. This host should probably be forcibly (-f) deleted.

host nodename is already in the set

An attempt was made to add a host nodename which already exists within the diskset.

host nodename is modifying set - try later or restart rpc.metad

Either an attempt was made to perform an operation on a diskset at the same time as someone else, or a previous operation dropped core and the rpc.metad daemon should be restarted on host nodename.

host nodename is not in the set

An attempt was made to delete the host nodename from a diskset which does not contain the host.

host nodename is specified more than once

The same host (nodename) was specifed more than once in the command line.

host name nodename is too long

The name used for the host nodename is longer than DiskSuite accepts.

hotspare doesn't exist

An attempt was made to perform an operation on the hot spare dev and the specified hot spare does not exist.

hotspare in use

An attempt was made to perform an operation on the hot spare dev and the specified hot spare is in use.

hotspare isn't broken, can't enable

An attempt was made to enable a hot spare that is not in the broken state.

hotspare database create failure

An attempt to create a hot spare record in the metadevice state database failed. Run metadb -i to determine the cause of the failure.

hotspare pool database create failure

An attempt to create a hot spare pool record in the metadevice state database failed. Run metadb -i to determine the cause of the failure.

hotspare pool is busy

An attempt was made to delete the hot spare pool hspnnn before removing all the hot spares associated with the specified hot spare pool.

hotspare pool is referenced

An attempt was made to delete the hot spare pool, hspnnn, that is associated with a metadevice.

hotspare pool in use

An attempt was made to metaclear(1M) a hotspare pool without first removing its association with metadevices.

hotspare pool is already setup

An attempt was made to create a hot spare pool which already exists.

illegal option

An attempt was made to use an option which is not valid in the context of the specified metadevice or command.

in Last Erred state, errored components must be replaced

An attempt was made to replace or enable a component of a mirror in the "Last Erred" state when other components are in the "Erred" state. You must first replace or enable all of the components in the "Erred" state.

invalid RAID configuration

An invalid RAID device configuration entry was supplied to metainit, either from the command line or via the md.tab file.

invalid argument

An attempt was made to use an argument which is not valid in the context of the specified metadevice or command.

invalid column count

An invalid RAID configuration entry was supplied to metainit, either from the command line or via the md.tab file. Specifically, an invalid argument was provided with the -o option.

invalid interlace

An unsupported interlace value follows the -i option on a metadevice configuration line. The -i specifies the interlace size. The interlace size is a number (8, 16, 32) followed by either k for kilobytes, m for megabytes, or b for blocks. The units can be either uppercase or lowercase. This message will also appear if the interlace size specified is greater than 100 Mbytes.

invalid mirror configuration

An invalid mirror configuration entry was supplied to metainit, either from the command line or via the md.tab file.

invalid pass number

An attempt was made to use a pass number for a mirror that is within the 0 - 9 range.

invalid stripe configuration

An invalid stripe configuration entry was supplied to metainit, either from the command line or via the md.tab file.

invalid trans configuration

An invalid trans configuration entry was supplied to metainit, either from the command line or via the md.tab file.

invalid write option

An attempt was made to change the write option on a mirror using an invalid option. The legal strings are "serial" and "parallel."

invalid hotspare pool

The metadevice configuration entry in the md.tab file has a -h hspnnn and a metainit has not been performed on the hot spare pool.

invalid read option

The user has specified both the -r and -g options on the same metamirror.

invalid unit

The metadevice (submirror) passed to metattach is already a submirror. The metadevice may already be a submirror for another metamirror.

is a metadevice

The device dev being used is a metadevice and it should be a physical component.

is mounted on

The device dev in the metadevice configuration has a file system mounted on it.

hostname is not a nodename, but a network name

An attempt was made to add a host to a diskset without using the nodename found in the /etc/nodename file.

is swapped on

The device in the metadevice configuration is currently being used as a swap device.

maximum number of nodenames exceeded

An attempt was made to add more nodenames than DiskSuite allows in a diskset.

maxtransfer is too small

An attempt was made to add a component to a RAID device whose maxtransfer is smaller than the other components in the RAID device.

metadevice in use

An attempt was made to metaclear(1M) a submirror without first running metaclear on the metamirror in which it is contained.

metadevice is open

The metadevice (submirror) passed to metattach is already open (in-use) as a metadevice.

num1 metadevice database replicas is too many;the maximum is num2

An attempt was made to add more databases (num1) than the maximum allowed (num2).

metadevice database has too few replicas, can't create new records

An attempt to create a metadevice record in the metadevice state database failed. Run metadb -a to add more database replicas.

metadevice database is full, can't create new records

An attempt to create a metadevice record in the metadevice state database failed. Run metadb -a (and -s) to add larger database replicas. Then delete the smaller replicas.

metadevice database replica exists on device

An attempt was made to use a component (ie. for a hot spare) which contains a database replica.

mirror has maximum number of submirrors

An attempt was made to attach more than the supported number of submirrors. The maximum supported number of submirrors is three.

must be owner of the set for this command

An attempt was made to perform an operation on a diskset or a shared metadevice on a host which is not the owner of the diskset.

must have at least 2 databases (-f overrides)

An attempt was made to delete database replicas, reducing the number of database replicas to a number less than two. To override this restriction, use the -f option.

must replace errored component first

An attempt was made to replace or enable a component of a mirror in the "Last Erred" state when other components are in the "Erred" state. You must first replace or enable all of the components in the "Erred" state.

no available set numbers

An attempt was made to create more disksets than DiskSuite allows.

no hotspare pools found

An metahs operation was attempted using the "all" argument when no hot spare pools meet the criteria for the operation.

no metadevice database replica on device

An attempt was made to delete non-existent database replicas.

no such set

An attempt was made to perform an operation on a diskset or a shared metadevice using a non-existent set name.

nodename of host nodename creating the set must be included

An attempt was made to create a diskset on the local host without adding the name of the local host to the diskset.

not a disk device

The component name specified is not a disk device name. For example, a CD-ROM device doesn't have the characteristics of a disk device.

not enough components specified

An invalid stripe configuration entry was supplied to metainit, either from the command line or via the md.tab file.

not enough stripes specified

Invalid stripe configuration entry was supplied to metainit, either from the command line or via the md.tab file.

not enough submirrors specified

Invalid mirror configuration entry was supplied to metainit, either from the command line or via the md.tab file.

not in local set

An attempt was made to create a local metadevice or local hot spare pool with a component whose drive is contained in a shared diskset.

not a metadevice

The specified device is not a metadevice. DiskSuite expected a metadevice name.

only slice 7 is usable for a diskset database replica

An attempt was made to add a database replica for a shared diskset on a component other than Slice 7.

only the current owner nodename may operate on this set

An attempt was made to perform an operation on a diskset or a shared metadevice on a host which is not the owner of the diskset.

only valid action is metaclear

The initialization of a RAID device has failed. Use the metaclear command to clear the RAID device.

operation would result in no readable submirrors

An operation was attempted on a component or submirror that contains the only copy of the data. The other submirrors have erred components. If this operation were allowed, the mirror would be unusable.

operation requires -f (force) flag

Due to the components within the RAID device being in the "Maintenance" or "Last Erred" state, the force flag (-f) is required to complete the operation.

overlaps with device in metadevice

Overlapping slices are not allowed in metadevices or hot spare pools.

replace failure, new component is too small

An attempt to use metareplace failed because the new component is too small to replace the old component.

reserved by another host

An attempt was made to add a currently reserved drive to a diskset.

resync in progress

The mirror operation failed because a resync is being performed on the specified metamirror. Retry this operation when the resync is finished.

set setname is out of date - cleaning up - take failed

The diskset setname is out of data with respect to the other host's view. This error should occur only after one-host administration.

set lock failed - bad key

Either an operation on a diskset is currently being performed by another host, or an operation on a diskset was aborted. In the latter case, rpc.metad should be killed.

set name contains invalid characters

An attempt was made to use illegal characters to name a diskset.

set name is in-use or invalid on host nodename

The diskset name selected is already in use on host nodename or contains characters not considered valid in a diskset name.

set name is too long

An attempt was made to create a diskset using more characters in the diskset name than DiskSuite will accept.

set unlock failed - bad key

The diskset is locked and the user does not have the key. This may require killing rpc.metad.

side information missing for host nodename

The diskset is incomplete. Kill rpc.metad on all hosts and then retry the operation.

slice 7 is not usable as a metadevice component

An attempt was made to use Slice 7 in a shared metadevice or shared hot spare pool. Slice 7 is reserved for database replicas only.

submirror too small to attach

The metadevice passed to metattach is smaller than the metamirror to which it is being attached.

stale databases

The user attempted to modify the configuration of a metadevice when at least half the metadevice state database replicas were not accessible.

syntax error

An invalid metadevice confguration entry was provided to metainit from the command line or via the md.tab file.

there are no existing databases

To create any metadevices or hot spare pools, database replicas must exist. See metadb(1M) for information on the creation of database replicas.

unable to delete set, it still has drives

An attempt was made to delete the last remaining host from a diskset while drives still exist in the diskset.

unit already set up

The user requested that a metadevice dnum be initialized when dnum is already set up.

unit is not a concat/stripe

An attempt was made to perform a concat/stripe specific operation on a metadevice that is not a concat/stripe.

unit is not a mirror

An attempt was made to perform a mirror specific operation on a metadevice that is not a mirror.

unit is not a RAID

An attempt was made to perform a RAID specific operation on a metadevice that is not a RAID device.

unit is not a trans

An attempt was made to perform a metatrans specific operation on a metadevice that is not a metatrans device.

unit not found

An attempt was made to perform an operation on a non-existent metadevice.

unit not set up

An attempt was made to perform an operation on a non-existent metadevice.

waiting on /tmp/.mdlock

Some other metadevice utility is currently in progress and the lock cannot be accessed at this time. DiskSuite utilities are serialized using the /tmp/.mdlock file as a lock. If you determine that there are no other utilities currently running, you may want to remove this lock file.

Log Messages

The log messages displayed by DiskSuite are listed in alphabetical order below. Each message is always preceded with "md:" The variables in these messages indicate the following:

• dev is a device name.
• dnum is a metadevice name.
• num is a number.
• state is a metatrans device state
• trans is either "logging" or "master"

---

Note - When the initial portion of a message begins with a variable, the message is alphabetized by the first word following the variable.

---

Notice Log Messages

Could not load misc /dev

The named misc module is not loadable. It is possibly missing, or something else has been copied over it.

db: Parsing error on 'dev'

The set command in /etc/system for the mddb.bootlist<number> is not in the correct format. Run metadb -p to place the correct set commands into the /etc/system file.

dnum: Hotspared device dev with dev

The first device name listed has been hot spare replaced with the second device name listed.

dnum: Hotspared device dev(num, num) with dev(num, num)

The first device number listed has been hot spare replaced with the second device number listed.

dnum: no mem for property dev

Memory could not be allocated in the prop_op entry point.

Warning Log Messages

dnum: Cannot load dev driver

The underlying named driver module is not loadable (for example, sd, id, xy, or a third-party driver). This could indicate that the driver module has been removed.

Open error of hotspare dev Open error of hotspare dev(num, num)

The named hot spare is not openable, or the underlying driver is not loadable.

dnum: read error on dev dnum: write error on dev

A read or write error has occurred on the specified metadevice at the specified device name. This happens if any read or write errors occur on a metadevice.

dnum: read error on dev(num, num) dnum: write error on dev(num, num)

A read or write error has occurred on the specified metadevice at the specified device number. This happens if any read or write errors occur on a metadevice.

dnum: read error on dnum dnum: write error on dnum

A read or write error has occurred on the specified metadevice at the specified device number. This happens if any read or write errors occur on a metadevice.

State database commit failed State database delete failed

These messages occur when there have been device errors on components where the state database replicas reside. These errors only occur when more than half of the replicas have had errors returned to them. For example, if you have three components with state database replicas and two of the components report errors, than these errors may occur. The state database commit or delete is retried periodically. If the replica is added, the commit or delete will finish and the system will be operational. Otherwise, the system will time out and panic.

State database is stale

This message occurs when there are not enough usable replicas for the state database to be able to update records in the database. All accesses to the metadevice driver will fail. To fix this problem, add more replicas or delete unaccessible replicas.

trans device: read error on dnum trans device: write error on dnum

A read or write error has occurred on the specified logging or master device at the specified metadevice. This happens if any read or write errors occur on a logging or master device.

trans device: read error on dev trans device: write error on dev

A read or write error has occurred on the specified logging or master device at the specified device name. This happens if any read or write errors occur on a logging or master device.

trans device: read error on dev(num, num) trans device: write error on dev(num, num)

A read or write error has occurred on the specified logging or master device at the specified device number. This happens if any read or write errors occur on a logging or master device.

logging device: dnum changed state to state logging device: dev changed state to state logging device: dev(num, num) changed state to state

The logging device and its associated master device(s) have changed to the specified state(s).

Panic Log Messages

State database problem

A failed metadevice state database commit or deletion has been retried the default 100 times.

dnum: Unknown close type dnum: Unknown open type

A metadevice is being opened/closed with an unknown open type (OTYP).