Glossary

asynchronous queue

A local area of disk or memory used to store writes that are to be replicated to a remote site. After the writes have been put into the queue, the write is acknowledged to the application, and the writes are forwarded to the remote site at a later time, as the network capabilities permit.

asynchronous replication

A synchronous replication confirms to the originating host that the primary I/O transaction is complete before updating the remote image. That is, completion of the I/O transaction is acknowledged to the host when the local write operation is finished and the remote write operation has been queued. Deferring the secondary copy removes the long distance propagation delays from the I/O response time.

auto synchronization

With the auto synchronization option enabled on the primary host, the synchronization daemon (autosyncd) attempts to resynchronize volume sets if the system reboots or a link failure occurs.

blocking

(asynchronous queue) In blocking mode, if the asynchronous queue fills, all future writes are delayed until the queue drains enough to allow for a write to occur. Blocking mode, which is the default Asynchronous running option, ensures write ordering of the packets to the secondary site. If the asynchronous queue fills with the blocking option set, response time to the application may be impacted.

configuration location

Location where the Sun StorageTek Availability Suite software stores configuration information about all enabled volumes used by the software.

consistency group

A consistency group is a group of remote volumes that share a single asynchronous queue to maintain w rite ordering.

dsstat

A tool from the Sun StorageTek Availability Suite tool set that can be used to display kernel statistics from the Remote Mirror and Point-in-Time snapshot products.

firewall

A computer that acts as an interface between two networks and regulates traffic between those networks for the purpose of protecting the internal network from electronic attacks originating from the external network.

forward resynchronization

See Update synchronization.

full synchronization

Full synchronization performs a complete volume-to-volume copy, which is the most time-consuming of the synchronization operations. In most cases, a secondary volume is synchronized from its source primary volume. However, restoration of a failed primary disk might require reverse synchronization, using the surviving Remote Mirror as the source.

hwm

See High water mark

high water mark

The high water mark is the largest amount of the asynchronous queue that has been used.

lazy clear

The practice of clearing the bit in core, but not writing the bitmap block back to disk until either another bit is set, or the in-core copy is reclaimed. This is safe because the change will simply be retransmitted after a system failure.

logging

Mode where a bitmap tracks writes to a disk, rather than a running log of each I/O event. This method tracks disk updates that have not been remotely copied while the remote service is interrupted or impaired. The blocks that no longer match their remote sets are identified for each source volume. The software uses this log to re-establish a Remote Mirror through an optimized update synchronization rather than a complete volume-to-volume copy.

non-blocking

(asynchronous queue) In non-blocking mode, if the asynchronous queue fills, the Remote Mirror software goes into scoreboarding mode and the contents of the queue are discarded. Non-blocking mode, does not ensure write ordering of the packets to the secondary site, but it assures that response time to the application will not be impacted if the asynchronous queue fills.

primary or local: host or volume

The system or volume on which the host application is principally dependent. For example, this is where the production database is being accessed. This data is to be replicated to the secondary by the software.

replication

Once a volume set has been initially synchronized, the software ensures that the primary and secondary volumes contain the same data on an ongoing basis. Replication is driven by user-layer application write operations; replication is an ongoing process.

reverse synchronization

An operation used during recovery rehearsals. Logging keeps track of test updates applied to the secondary system during the rehearsal. When the primary is restored, the test updates are overwritten with the blocks from the primary image, restoring matching remote sets.

secondary or remote: host or volume

The remote counterpart of the primary, where data copies are written to and read from. Remote copies are transmitted without host intervention between peer servers. A server might act as primary storage for some volumes and secondary (remote) storage for others.

synchronization

The process of establishing an identical copy of a source disk onto a target disk as a precondition to the software mirroring.

synchronous replication

Synchronous replication is limited to short distances (tens of kilometers) because of the detrimental effect of propagation delay on I/O response times.

TCP buffer

The TCP buffer size is the number of bytes that the transfer control protocol allows to be transferred before it waits for an acknowledgment.

update synchronization

Update synchronization copies only those disk blocks identified by logging, reducing the time to restore remotely mirrored sets.

volume set file

A text file containing information about specific volume sets. This text file is not the same as the configuration location, which contains information about all configured volume sets used by the Remote Mirror and Point-in-Time Copy software.