IBM Flex System V7000 Storage Node 4939H49 User Manual
Product codes
4939H49
8068ch09-Copy Servies.fm
Draft Document for Review January 29, 2013 12:52 pm
398
IBM Flex System V7000 Storage Node Introduction and Implementation Guide
Figure 9-68 The effect of packet size (in bytes) versus the link size
Before you proceed, take a quick look at the second largest component of your
round-trip-time, that is, serialization delay. Serialization delay is the amount of time required to
move a packet of data of a specific size across a network link of a given bandwidth. This delay
is based on a simple concept: the time required to move a specific amount of data decreases
as the data transmission rate increases.
round-trip-time, that is, serialization delay. Serialization delay is the amount of time required to
move a packet of data of a specific size across a network link of a given bandwidth. This delay
is based on a simple concept: the time required to move a specific amount of data decreases
as the data transmission rate increases.
In Figure 9-68, there are orders of magnitude of difference between the different link
bandwidths. It is easy to see how 1920 errors can arise when your bandwidth is insufficient
and why you should never use a TCP/IP
bandwidths. It is easy to see how 1920 errors can arise when your bandwidth is insufficient
and why you should never use a TCP/IP
ping
to measure RTT for FCIP traffic.
Figure 9-68 compares the amount of time in microseconds required to transmit a packet
across network links of varying bandwidth capacity. Three packet sizes are used:
across network links of varying bandwidth capacity. Three packet sizes are used:
��
64 bytes: The size of the common ping packet
��
1500 bytes: The size of the standard TCP/IP packet
��
2148 bytes: The size of a Fibre Channel frame
Remember that your path MTU affects the delay incurred in getting a packet from one location
to another, when it causes fragmentation or is too large and causes too many retransmits
when a packet is lost.
to another, when it causes fragmentation or is too large and causes too many retransmits
when a packet is lost.
9.4.2 1720 error
The 1720 error (event ID 050020) is the other primary error code of Remote Copy. Because
the term “System Partnership” implies that all involved virtualization systems are partners,
they must be able to communicate with each other. When a partner on either side stops
communicating, you see a 1720 error appear in your error log. According to official
documentation, there are no likely field replaceable unit breakages or other causes.
the term “System Partnership” implies that all involved virtualization systems are partners,
they must be able to communicate with each other. When a partner on either side stops
communicating, you see a 1720 error appear in your error log. According to official
documentation, there are no likely field replaceable unit breakages or other causes.
In practice, the source of this error is most often a fabric problem or a problem the network
path between your partners. When you receive this error, if your fabric has more than 64 HBA
ports zoned, you should check your fabric configuration for zoning of more than one HBA port
path between your partners. When you receive this error, if your fabric has more than 64 HBA
ports zoned, you should check your fabric configuration for zoning of more than one HBA port