vxrlink

NAME

vxrlink - perform VERITAS Volume Manager operations on RLINKs

SYNOPSIS

DESCRIPTION

The vxrlink utility performs VERITAS Volume Manager (VxVM) operations on RLINKs (rlinks). vxrlink is specific to VERITAS Volume Replicator (VVR), which requires a valid license. The first operand is a keyword that determines the specific operation to perform.

Each operation can be applied to only one disk group at a time. Any rlink and rvg operands are used to determine a default local disk group, according to the standard disk group selection rules described in vxintro(1M). When an rvg\ is not specified, the RVG currently associated with the specified RLINK is used. A specific local disk group can be selected with -g diskgroup.

KEYWORDS

assoc

Associates an rlink with an rvg. Both the rlink record and the rvg record are modified. Alternatively the association could be specified during the vxmake of the RVG. Secondary RVGs (that is, an RVG that does not have the primary attribute set) have at most one rlink associated with them. Where as a primary RVG has one rlink for each secondary RVG associated with it with a maximum of 32.

att

Attaches an rlink to an rvg. The rlink must already be associated with the rvg before the attach operation.

When a primary rlink is successfully attached, VVR sets up a connection between the primary and secondary hosts and replicates the writes received by the primary data volumes to the corresponding secondary data volumes. Before attaching a primary rlink ensure that the secondary RLINKs are attached, the secondary Storage Replicator Log(SRL) is associated with the secondary RVG, and all of the secondary data volumes are associated with the secondary RVG. For the attach to succeed on the primary, the remote_host, remote_dg, and remote_rlink attributes must be set on both the primary and the secondary RLINKs. The attributes can be set, while creating the RLINKs, using the vxmake or vxedit set command. The attach fails if the primary RVG does not have an SRL associated with it. The exception to this rule is when the primary RVG is in passthru error mode and does not need an SRL associated with it. In this case, the application writes are passed directly to the data volumes and replication is not taking place.

An attach on the primary requires that one of the options -a, -c checkpoint, or -f is specified.

An attach of a secondary RLINK does not require any options because the attach only indicate the readiness of the RLINK to be used for replication.

checkdelete

Deletes the specified Secondary checkpoint.

By default, VVR allows the creation of a maximum of 46 checkpoints. When the number of checkpoints exceeds this number, VVR prompts you to select which of the checkpoints should be deleted.

cplist

Displays information about all existing checkpoints associated with the given rlink, including their name, size, and percentage of SRL used. If SRL usage reaches 100%, the checkpoint overflows and becomes unusable, as indicated by the output message "<Checkpoint overflowed>." Note that only secondary checkpoints (created using vxrlink -c checkpoint pause on the secondary) are associated with an rlink. Primary checkpoints (created using vxrvg -c checkpoint checkstart on the primary) are associated with the RVG, and can be listed using vxrvg cplist.

det

Detaches an rlink from an rvg. This makes the RLINK STALE.

dis

Dissociates an rlink from the rvg with which it is associated. This can not be executed if the RLINK is currently attached.

pause

Pauses updates to the rlink until vxrlink resume is executed. New updates are logged while the RLINK is paused, and are applied once the RLINK is resumed.

On the primary, if the Data Change Maps(DCMs) are being replayed on the RLINK to be paused, the replay pauses until the RLINK is resumed. If more than one RLINK is involved in the replay, the replay pauses for all RLINKs involved. Resuming the paused RLINK allows the replay to resume for all RLINKs. Detaching the paused RLINK allows the replay to resume for the remaining attached RLINKs.

On the secondary a -c checkpoint option is valid and can be used to mark a point at which a backup of the secondary has been taken.

recover

Recovers rlink after rebooting machine.

restore

Restores the rlink from FAIL state to ACTIVE. Valid only for secondary RLINKs. Used when restoring data volumes at a secondary site from online backup data maintained at the secondary site (as opposed to restoring secondary data volumes using data copied from the primary site).

resume

Resumes updates to an rlink in the PAUSE state. After updates are resumed, all updates that were logged while the RLINK was paused are written to the secondary.

stats

Displays network statistics for the given rlink. The displayed values are cumulative, except when used with the -i interval option. When used with -i interval option, the first line of statistics provide cumulative values and subsequent lines show the change during the last interval seconds.

The following values are displayed under the heading Messages:

#

Number of messages transmitted.

Blocks

Number of 1024-byte blocks transmitted.

RT

Average round-trip time per message.

The following values are displayed under the heading Errors:

Timeout

Number of timeouts errors. A timeout error occurs when an acknowledgement for a message is not received from the remote host within the timeout period.

Stream

Number of stream errors. Stream errors occur when the RLINK attempts to send messages faster than the network can handle.

Memory

Number of errors because of insufficient buffer space on the Secondary.

The heading Flow Control displays values that indicate how fast the RLINK is trying to send messages. If VVR is using the UDP transport protocol, the values Delays, NW Bytes and NW Delay are displayed. If VVR is using the TCP transport protocol, only the Delays value is displayed.

Delays

Cumulative time the RLINK had to wait while it was flow controlled. The value is displayed in one-hundredth of a second.

NW Bytes

Number of bytes that can be transmitted without flow controlling and introducing any intervening wait. If an RLINK does not experience network errors, VVR steadily increases the NW Bytes to permit more data to be transmitted. If an RLINK experiences network error, VVR tries to flow control by reducing this number. The minumum value is 5000 bytes.

NW Delay

The amount of wait introduced after transmission of NW Bytes of a message. If NW Bytes is at 5000 bytes and the RLINK continues to experience errors, VVR increases NW Delay. The value is displayed in one-hundredth of a second. The maximum value is 100 (1 second). A steady high value of NW Delay usually indicates a network problem.

If VVR is using the UDP transport protocol, the current value of Timeout is displayed:

Timeout

Message timeout value in milli-seconds. The value is computed dynamically. If an acknowledgement for a message is not received from the remote host within this time period, the message is considered lost and is retransmitted.

When used with the -e option, the stats command displays extended network statistics. The displayed values are cumulative unless an interval is specified using the -i option. In this case, cumulative values are only shown on the first line of statistics, and subsequent lines show the change during the previous interval.

The output of the command with the -e option is specified is displayed under two headings: Messages and Errors. The Messages heading displays the following information:

Number of blocks sent

The number of 512-byte blocks that have been transmitted. (If the -e option is not specified, the displayed Blocks value only indicates the number of blocks that have been acknowledged.)

The Errors heading has nine fields that indicate the nature of the error that has been encountered. Three of these headings are similar to those shown by the command if the -e option is not specified. The following errors may be shown:

No memory available

There is not enough system kernel memory to process the message.

No message slots available

There is not enough memory to store the packets of a message that have arrived out of sequence. (If a packet arrives out of sequence, it must be stored in a message buffer until all the related out-of-sequence packets arrive, and can be assembled.)

No memory available in nmcom pool on Secondary

The buffer space (whose size is determined by the value of VVR tunable vol_max_nmpool_sz) is full, and cannot store any new messages that arrive at the Secondary.

Timeout errors

Displays the number of times that the Primary timed out while waiting for an acknowledgement from the secondary.

Missing packet errors

Displays the number of times that the last packet of a message was received before one or more packets of the same message had been received.

Missing message errors

Displays the number of times that messages have arrived out of sequence.

Stream errors

Displays the number of times that the RLINK has attempted to send messages faster than the network can handle.

Checksum errors

Displays the number of data checksum errors. (Every time a packet is received at the secondary, VVR performs a checksum to ensure that the packet data is the same as that sent by the Primary.)

Unable to deliver due to transaction errors

Displays the number of times that packets could not be delivered to the secondary because of transaction errors. (If the secondary is busy with some kernel operations when a packet arrives at the secondary, the packet may not be delivered until the transaction is complete.)

status

Only valid on the primary, status displays the state of the secondary image represented by the given rlink. If the rlink is not up-to-date, the output shows the number of updates and blocks by which the image is behind, and the percentage of the SRL being used by this rlink. The -i interval option allows the status command to be executed at the given interval.

The -T option (timestamp) displays the status of the Primary showing the amount of time (as hours:minutes:seconds) by which the given RLINK is lagging.

updates

Only valid on the secondary, updates displays the update ID of the latest update received by the secondary, as well as number of updates primary is ahead of secondary. This information can be used to determine the most up-to-date secondary RVG. The update information is invalid if the secondary has not received any updates from the primary. The update information is unavailable if the secondary has been rebooted after the primary has gone down.

The -T option (timestamp) displays the amount of time by which the secondary is lagging (as Mon DOM hours:minutes:seconds where Mon is the locale-specific abbreviation of the month, and DOM is the day number in the month).

verify

Displays configuration information for the rlink. Configuration update messages are sent to each secondary RLINK when appropriate. Any configuration errors detected at a secondary are logged. This command displays the configuration status of the given RLINK. If an rvg is specified instead of an RLINK, configuration status information is displayed for each RLINK associated with the RVG. If an RVG is specified with the -r option, it is used only to help determine a default local disk group, according to the standard disk group selection rules described in vxintro(1M).

OPTIONS

-a

Use with the att command on the primary to attach a group of RLINKs for automatic synchronization(autosync). The autosync operation fails if all the data volumes do not have associated DCMs or if any of the RLINKs are unable to connect to the respective secondary hosts. The -a option cannot be used on the secondary or with the -f or -c option.

-c checkpoint

Use with the att command on the primary to attach an rlink which is consistent up to the point indicated by the checkpoint string. The checkpoint indicates, to VVR, the writes that must be replayed from the primary host to the seconday hosts, to bring the secondary hosts up to date. The -c option is also required with the restore command to specify the checkpoint to be used for restore. Similarly, the -c option is required with the checkdelete command to specify a checkpoint that is to be deleted.

-e

Displays extended statistics for errors reported by the RLINK. The output includes six additional columns that provide diagnostic information on the error conditions. This option can be used in conjunction with the -i and -t options. In a clustered environment extended statistics will be printed only on a node which is the logowner and is running the CVM protocol version 60 or higher.

-f

Use with the att command on the primary to force an attach of an rlink when the secondary data volumes are completely synchronized with the primary data volumes.

Use the -f option with care because it may result in loss of data if used incorrectly. Use the -f option to attach the rlink only if you are sure that the rlink is up to date.

-g diskgroup

Specifies the local diskgroup for the operation. The diskgroup can be specified either by name or by disk group ID. See vxdg(1M) for more information on disk groups.

-i interval

Executes the stats or status command every interval seconds.

-r rvg

Specifies the name of the RVG with which the rlink is associated. If not supplied, the RLINK record is examined to retrieve the name of the associated RVG.

-t timestamp_freq

Specifies how frequently timestamps are to be output, in terms of the number of lines. For example, -t 1 prints a timestamp on each line, and -t 5 prints a timestamp on every fifth line. This option is only valid for the stats and status commands when the -i option is also specified.

-T

For the vxrlink status command, displays the status of the Secondary image represented by the given RLINK together with a timestamp that indicates the amount of time by which the RLINK is lagging.

For the vxrlink updates command, displays the amount of time by which the Secondary is lagging.

-w

Forces a secondary rlink into the FAIL state. Used only in special circumstances (secondary online backup). The rlink is marked inconsistent.

EXIT CODES

The vxrlink utility exits with a non-zero status if the attempted operation fails. A non-zero exit code is not a complete indicator of the problems encountered, but rather denotes the first condition that prevented further execution of the utility.

See vxintro(1M) for a list of standard exit codes.

SEE ALSO

vxdg (1M), vxedit (1M), vxintro (1M), vxmake (1M), vxplex (1M), vxprint (1M), vxrvg (1M), vxvol (1M), vradmin (1M)