HP StorageWorks 2/16V HP StorageWorks Fabric OS 5.3.1b release notes (5697-764 - Page 23

Fabric OS 5.3.0b, Fabric OS 5.3.0c

Page 23 highlights

Fabric OS 5.3.0b fixes Table 5 lists defects closed in the Fabric OS 5.3.0b firmware release. Table 5 Fabric OS 5.3.0b closed defects Closed defect summary Solution With FC Tape Pipelining enabled on a 400 MP Router/MP Router Blade FCIP connection, data written to tape using a block size of greater than 128 KB will not be readable by some applications, causing backup to be unreadable. Backup application could be configured to use less than 128 KB block size to avoid the problem, but this is now fixed in 5.3.0b. In rare cases, CRC errors can occur on Brocade 4Gb SAN Switch for Fabric OS 5.3.0b addresses this HP c-Class BladeSystem internal ports when operating at 4Gb/s. issue. When supportsave is run in a continuous loop, certain types of files are not detected which periodically generates the message: SupportSave completed sh: /tmp/supportSave_files/CONSOLE0: No such file or directory. File name can vary. Fixed in Fabric OS 5.3.0b. When the SFP/XFP power and voltage readings are above or below the range, an excessive number of alert messages were reported. Fixed in Fabric OS 5.3.0b. Third-party application will not work with the 400 MP Router Fast Write application (sends FCP_RSP without sequence initiative). Fixed in Fabric OS 5.3.0b. On the B-Series MP Router Blade or 400 MP Router, the proxy PID is improperly converted to 000000 for an ACC to a REC extended link request when an EX_Port is configured for interop mode. As a result, Tape backup may not work with devices in McDATA edge fabrics across the Router due to an improper response to a REC extended link request to a tape device. Fixed in Fabric OS 5.3.0b. Fabric OS 5.3.0c fixes Table 6 lists defects closed in the Fabric OS 5.3.0c firmware release. Table 6 Fabric OS 5.3.0c closed defects Closed defect summary Solution I/O Ops failure due to incorrect CUP Port PDCM bit for port 256 after firmwareupgrade from FOS 5.1.x to FOS 5.2.x/5.3.x in FICON Fixed in Fabric OS 5.3.0c. environment. With McDATA edge switches in the fabric, 400 MP Router or FR4-18i router port receives periodic RRQ & ABTS frames due to the McDATA switch sending unsupported CT command. Routers don't respond, which causes McDATA switch to retry indefinitely. Fixed in Fabric OS 5.3.0c to reject the unsupported CT command. After switchdisable & switchenable, all 400 MP Routers or FR4-18i routers, proxydevices and nameserver entries no longer match what was previously recorded. "fcrrproxydevshow -a" shows 0 devices and translate domains are missing from the edge fabric. This is due to a timing issue while in open fabric interop mode with i10k switch when 2nd IFL is connected to the same switch in the fabric as the 1st IFL, and 1st IFL is fully up. Fixed in Fabric OS 5.3.0c. Hosts lose paths to targets due to FCR daemon failing to send GA_NXT response from 400 MP Router to host connecting to McDATA switch in open fabric mode. Caused by incorrect domain offset in its address translation. Fixed in Fabric OS 5.3.0c. HP StorageWorks Fabric OS 5.3.1b release notes 23

  • 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
  • 27
  • 28

Fabric OS 5.3.0b
xes
Table 5
lists defects closed in the Fabric OS 5.3.0b
rmware release.
Table 5 Fabric OS 5.3.0b closed defects
Closed defect summary
Solution
With FC Tape Pipelining enabled on a 400 MP Router/MP Router
Blade FCIP connection, data written to tape using a block size of
greater than 128 KB will not be readable by some applications,
causing backup to be unreadable.
Backup application could be
con
gured to use less than 128 KB
block size to avoid the problem, but
this is now
xed in 5.3.0b.
In rare cases, CRC errors can occur on Brocade 4Gb SAN Switch for
HP c-Class BladeSystem internal ports when operating at 4Gb/s.
Fabric OS 5.3.0b addresses this
issue.
When
supportsave
is run in a continuous loop, certain types of
les are not detected which periodically generates the message:
SupportSave completed sh:
/tmp/support-
Save_files/CONSOLE0:
No such file or directory.
File name can vary.
Fixed in Fabric OS 5.3.0b.
When the SFP/XFP power and voltage readings are above or below
the range, an excessive number of alert messages were reported.
Fixed in Fabric OS 5.3.0b.
Third-party application will not work with the 400 MP Router Fast
Write application (sends FCP_RSP without sequence initiative).
Fixed in Fabric OS 5.3.0b.
On the B-Series MP Router Blade or 400 MP Router, the proxy PID is
improperly converted to 000000 for an ACC to a REC extended link
request when an EX_Port is con
gured for interop mode. As a result,
Tape backup may not work with devices in McDATA edge fabrics
across the Router due to an improper response to a REC extended
link request to a tape device.
Fixed in Fabric OS 5.3.0b.
Fabric OS 5.3.0c
xes
Table 6
lists defects closed in the Fabric OS 5.3.0c
rmware release.
Table 6 Fabric OS 5.3.0c closed defects
Closed defect summary
Solution
I/O Ops failure due to incorrect CUP Port PDCM bit for port 256
after
rmwareupgrade from FOS 5.1.x to FOS 5.2.x/5.3.x in FICON
environment.
Fixed in Fabric OS 5.3.0c.
With McDATA edge switches in the fabric, 400 MP Router or FR4-18i
router port receives periodic RRQ & ABTS frames due to the McDATA
switch sending unsupported CT command. Routers don’t respond,
which causes McDATA switch to retry inde
nitely.
Fixed in Fabric OS 5.3.0c to reject
the unsupported
CT
command.
After
switchdisable
&
switchenable
, all 400 MP Routers
or FR4-18i routers, proxydevices and nameserver entries no longer
match what was previously recorded. "fcrrproxydevshow -a" shows 0
devices and translate domains are missing from the edge fabric. This
is due to a timing issue while in open fabric interop mode with i10k
switch when 2nd IFL is connected to the same switch in the fabric as
the 1st IFL, and 1st IFL is fully up.
Fixed in Fabric OS 5.3.0c.
Hosts lose paths to targets due to FCR daemon failing to send
GA_NXT response from 400 MP Router to host connecting to
McDATA switch in open fabric mode. Caused by incorrect domain
offset in its address translation.
Fixed in Fabric OS 5.3.0c.
HP StorageWorks Fabric OS 5.3.1b release notes
23