Dell VNX8000 White Paper: EMC Data Integrity on VNX

Dell VNX8000 Manual

Get Dell VNX8000 PDF manuals and user guides View all Dell VNX8000 manuals
Add to My Manuals
Save this manual to your list of manuals

Dell VNX8000 manual content summary:

  • Dell VNX8000 | White Paper: EMC Data Integrity on VNX - Page 1
    White Paper DATA INTEGRITY ON VNX Abstract This white paper explains the data integrity features on EMC® VNX™. March 2016
  • Dell VNX8000 | White Paper: EMC Data Integrity on VNX - Page 2
    Copyright © 2016 EMC Corporation. All Rights Reserved. EMC believes the information in this publication is accurate as of its publication date. The information is subject to change without notice. The information in this publication is provided "as is." EMC Corporation makes no representations or
  • Dell VNX8000 | White Paper: EMC Data Integrity on VNX - Page 3
    ...5 Protocols ...5 Ethernet Frame Check Sequence 5 Internet Protocol...6 Transmission Control and User Datagram Protocols 7 Fibre Channel Protocol ...7 Services ...8 Common Anti-Virus Agent 8 File-Level Retention ...8 MirrorView...9 VNX Replicator ...10 File System Check ...11 Mirrored SP
  • Dell VNX8000 | White Paper: EMC Data Integrity on VNX - Page 4
    Executive summary This white paper explains the methods used by EMC® VNX™ to ensure the end-toend integrity of data. Audience This white paper is intended for users interested in the technology used to ensure the safety and validity of their data. Data Integrity on VNX 4
  • Dell VNX8000 | White Paper: EMC Data Integrity on VNX - Page 5
    Introduction Data integrity refers to maintaining accuracy, completeness, and consistency of information throughout its entire life cycle. If the data is not valid, any analysis of that data will also be invalid. Data is at risk for corruption at any time, while in transit or at rest. EMC VNX
  • Dell VNX8000 | White Paper: EMC Data Integrity on VNX - Page 6
    Figure 2. IP Packet Header Figure 2 shows the contents of an IP packet header, which consist of: • 4-bit IP version • 4-bit header length • 8-bit service type • 16-bit total length • 16-bit identification tag • 3-bit fragment flags • 13-bit fragment offset • 8-bit time to live (TTL) • 8-bit protocol
  • Dell VNX8000 | White Paper: EMC Data Integrity on VNX - Page 7
    router whereas the data never changes. This decreases the processing time, because only the parts that may change are recalculated instead of the entire packet. The checksum is calculated by dividing the header into 16-bit words and summing them with a one's complement calculation. A one's
  • Dell VNX8000 | White Paper: EMC Data Integrity on VNX - Page 8
    not match, the frame is assumed to be corrupt and can be dropped. Services Common Anti-Virus Agent Data integrity can also be compromised by malware, such as refer to the Using VNX Event Enabler document available on EMC Online Support. File-Level Retention File-Level Retention (FLR) can be used to
  • Dell VNX8000 | White Paper: EMC Data Integrity on VNX - Page 9
    target individual files for deletion. For more information on FLR, refer to the EMC VNX File-Level Retention White Paper available on EMC Online Support. MirrorView It is also important to protect data that is being replicated for purposes such as disaster recovery (DR). This ensures the data on
  • Dell VNX8000 | White Paper: EMC Data Integrity on VNX - Page 10
    the size of the mirror images. When the secondary LUN returns to service, the secondary image must be synchronized with the primary image. This is to the MirrorView Knowledgebook: Releases 30-32 available on EMC Online Support. VNX Replicator VNX Replicator is an IP-based asynchronous tool that
  • Dell VNX8000 | White Paper: EMC Data Integrity on VNX - Page 11
    Replicator, refer to the Using VNX Replicator document available on EMC Online Support. File System Check File System Check (FSCK) is a tool used the FSCK are saved in the server_log. A FSCK can also be initiated manually if corruption is suspected. If corruption is suspected on a production file
  • Dell VNX8000 | White Paper: EMC Data Integrity on VNX - Page 12
    Figure 4. VNX SP Cache Architecture The only way the cached data can be lost is if both SPs fail at exactly the same time. Cached data is protected as long as the surviving SP has enough time to detect the failure on the other SP. In the event of an emergency, such as a power outage or temperature
  • Dell VNX8000 | White Paper: EMC Data Integrity on VNX - Page 13
    . RAID uses disk striping, mirroring, parity calculations, or a combination of the three to save data across multiple drives. VNX supports RAID 0, 1, 1/0, 3, 5, and 6. • RAID 0-Striped. RAID 0 has no redundancy but provides additional performance. Data blocks are written to multiple drives
  • Dell VNX8000 | White Paper: EMC Data Integrity on VNX - Page 14
    Using RAID technology (other than RAID 0) allows data integrity to be preserved through disk failures. Media errors, such as bad sectors, can be corrected by RAID instead of resulting in data loss. Parity Shedding When a disk has failed but has not yet been replaced, power interruptions can cause
  • Dell VNX8000 | White Paper: EMC Data Integrity on VNX - Page 15
    VNX1 systems offer enhanced RAID parity protection with parity shedding to protect against these events. Figure 6 shows the same scenario with parity shedding. 1. This is a RAID 5 (4+1) stripe in a consistent state. There are four data sectors and a parity sector with a value of 12. Note, parity
  • Dell VNX8000 | White Paper: EMC Data Integrity on VNX - Page 16
    of parity shedding, VNX2 systems leverage write journaling instead. Just like parity shedding, write journaling is designed to solve the RAID Write Hole problem where a spliced write on a power failure to a drive in a degraded RG could cause data loss. Parity shedding is no longer used because write
  • Dell VNX8000 | White Paper: EMC Data Integrity on VNX - Page 17
    on the drive. After the write is successfully committed to the drive, the journal entry is invalidated and the write is marked as clean in cache (Multicore Cache). Figure 7. Write Journaling Uncommitted dirty Multicore Cache pages are protected by the vaulting mechanism. After recovering a RAID
  • Dell VNX8000 | White Paper: EMC Data Integrity on VNX - Page 18
    bytes of metadata is a proprietary data integrity feature of VNX that helps address gaps in basic RAID. The metadata can be used for error detection and correction of sector-level errors. The eight additional bytes consist of: • 2-byte shed/LBA stamp • 2-byte checksum • 2-byte write stamp • 2-byte
  • Dell VNX8000 | White Paper: EMC Data Integrity on VNX - Page 19
    calculated with the same parity algorithm that is used in RAID. This allows the parity data to be maintained even when there are regions with bad blocks. Figure 8 shows how shed stamps work for RAID-3 and RAID-5. 1. The starting condition is an existing LUN with a write pending to disk 0. All shed
  • Dell VNX8000 | White Paper: EMC Data Integrity on VNX - Page 20
    Checksum The checksum is a 16-bit CRC calculated by running an XOR operation on all 512 bytes of user data. This is generated by the array any time data is written and stored within the sector. The checksum is always attached to the data as it travels between the cache and disk. It is used to detect
  • Dell VNX8000 | White Paper: EMC Data Integrity on VNX - Page 21
    2. Data is written to disks 1 and 2, and their respective write stamp bits are toggled. 3. Power is suddenly lost before the parity drive can be updated, and disk 2 fails while the power is down. When power is restored, the data from disk 2 is requested to be read. Because disk 2 is not available,
  • Dell VNX8000 | White Paper: EMC Data Integrity on VNX - Page 22
    Figure 11. Time Stamps Figure 12 shows how time stamps work in a worst-case failure scenario. 1. The starting condition is an empty LUN. All the data and time stamps are set to 0. 2. A full stripe write occurs. Data is written to disks 1 and 2, and the calculated parity is written to disk 4. 3. The
  • Dell VNX8000 | White Paper: EMC Data Integrity on VNX - Page 23
    is because time stamps are only used for full stripe writes, and write stamps are only used for non-full stripe writes. SNiiFFER Media problems or contamination from microscopic particles during normal operations can sometimes affect sectors on a disk. When this occurs, the controller might still be
  • Dell VNX8000 | White Paper: EMC Data Integrity on VNX - Page 24
    the stripe's calculated parity and the stored parity. When this occurs, the SP will attempt to reconstruct the data using the stored parity. A manual background verify can be started when the data integrity of a LUN is suspected to be compromised. The time required to run a full background verify
  • Dell VNX8000 | White Paper: EMC Data Integrity on VNX - Page 25
    stored data to ensure it is still valid before it is read. This is accomplished by integrating data integrity functionality into advanced services, mirroring SP cache, parity shedding, using NVRAM, 520-byte sectoring, the SNiiFFER process, and background verify. VNX features additional measures on
  • Dell VNX8000 | White Paper: EMC Data Integrity on VNX - Page 26
    from http://hsi.web.cern.ch/HSI/fcs/spec/overview.htm White Papers Refer to the following white papers, available on EMC Online Support, for more information: • MCx - Multicore Everything • Using VNX Event Enabler • EMC VNX File-Level Retention White Paper • MirrorView Knowledgebook: Releases 30-32
  • 1
  • 2
  • 3
  • 4
  • 5
  • 6
  • 7
  • 8
  • 9
  • 10
  • 11
  • 12
  • 13
  • 14
  • 15
  • 16
  • 17
  • 18
  • 19
  • 20
  • 21
  • 22
  • 23
  • 24
  • 25
  • 26
White Paper
Abstract
This white paper explains the data integrity features on EMC® VNX™.
March 2016
DATA INTEGRITY ON VNX