4.2.3Mirror Sequencing

Once a mirror has been created and fully replicated, it enters the INSYNC state. The mirror volume is finally mounted, read-only, and declared to be of partition type NBD. Then, the mirror begins sequencing. It is at this point that the mirror buffer begins to be actively used. Each file system transaction written to the mirror buffer on the master system is sent, or sequenced, to the mirror system. The master system waits for an acknowledgement from the mirror system that the transaction has been successfully written to its own mirror buffer before discarding it.

The mirror is considered to be "in sync", and will stay in this state, as long as file system changes on the master volume do not outpace the transfer of those changes to the mirror system. In the INSYNC state, the progress percentage represents the percentage of available transaction storage space in the mirror buffer; this percentage may fluctuate depending on disk usage and network latency.

4.2.4Link Down and Idle Conditions

At any point after the connection between the master and mirror systems has been established in the INIT state, a loss of communication between the two systems will cause the mirror to enter the LINKDOWN state. Data transfer from master to mirror is halted until the connection is re-established.

A mirror can also be manually paused, effecting the same behavior as a LINKDOWN condition. Pausing a mirror puts it into the IDLE state, and is only possible from the INSYNC state. In the IDLE state, transaction sequencing from the master to mirror is suspended.

4.2.5Cracked and Broken Mirrors

If the rate of file system changes on the master volume exceeds the rate of transfer to the mirror volume, the percentage will drop until it reaches zero. Once there is no more room in the mirror buffer for new transactions, the oldest transactions still pending acknowledgement from the mirror system will start to be overwritten. At this point, the mirror is considered to be "cracked", and will enter the CRACKED state. From the CRACKED state, the mirror reverts to the OUTOFSYNC state, the mirror volume is unmounted, and a full re-sync is started.

It is possible for the mirror to crack both when the mirror is sequencing, in the INSYNC state, and when the mirror is replicating, in the REPLICATE/OUTOFSYNC states. If the mirror buffer is not large enough to store all file system changes made before it finishes syncing to the mirror, it will continue in an endless OUTOFSYNC- CRACKED-OUTOFSYNC cycle.

4-12Sun StorEdge 5310 NAS Troubleshooting Guide • December 2004

Page 311 Page 313
Page 312
Image 312
Sun Microsystems 5310 NAS manual Mirror Sequencing, Link Down and Idle Conditions, Cracked and Broken Mirrors
Page 311 Page 313

5310 NAS specifications

Sun Microsystems, known for its innovative computing solutions, launched the Sun 5310 Network Attached Storage (NAS) system, which marked a significant advancement in the realm of storage solutions in the early 2000s. The 5310 NAS was designed to provide high-performance, reliable, and scalable storage tailored for enterprise environments.

One of the standout features of the Sun 5310 NAS is its file-serving capabilities, which support multiple protocols, notably NFS (Network File System) and CIFS (Common Internet File System). This dual-protocol support allowed organizations to seamlessly integrate the NAS into diverse IT ecosystems, facilitating interoperability between UNIX, Linux, and Windows systems. The enhanced file-sharing capabilities made it an ideal solution for businesses with mixed operating environments.

The Sun 5310 NAS incorporates cutting-edge technologies to ensure high availability and data integrity. The system utilized a RAID (Redundant Array of Independent Disks) technology, providing various RAID levels to protect against data loss while optimizing performance. Additionally, the device featured hot-swappable drives, enabling maintenance and upgrades with minimal downtime, a crucial factor for business continuity.

Equipped with advanced management software, the Sun 5310 NAS offered users an intuitive interface for monitoring storage health and performance. This software included comprehensive reporting functionalities that allowed IT administrators to oversee usage patterns and capacity planning efficiently.

Scalability was another defining characteristic of the Sun 5310 NAS. The system could easily expand with additional storage modules, accommodating the growing needs of an organization without the necessity for complete system overhauls. This flexibility ensured that businesses could adapt their storage solutions to meet evolving data storage needs without incurring significant costs or disruptions.

In terms of performance, the Sun 5310 NAS featured high I/O throughput achieved through its robust hardware architecture and optimized file serving capabilities. This performance baseline was crucial for organizations that relied on heavy data workloads and required rapid access to information.

In summary, the Sun 5310 NAS from Sun Microsystems embodied a forward-thinking approach to network storage, blending reliability, scalability, and multi-protocol support. Its user-friendly management software, combined with the robustness of RAID technology and hot-swappable drives, made it a preferred choice for enterprises looking to streamline their storage infrastructure while safeguarding critical data.
Top Page Image Contents