HP 6400/8400 HP StorageWorks Enterprise Virtual Array Updating Product Softwar - Page 25

Evaluating your configuration, XCS 09534000 upgrade considerations, ample

Get HP 6400/8400 PDF manuals and user guides View all HP 6400/8400 manuals
Add to My Manuals
Save this manual to your list of manuals

Page 25 highlights

Evaluating your configuration The following factors can affect the success of an online upgrade. Consider these factors when determining if an online upgrade is appropriate for your environment. • Queuing host I/O-After the upgrade, the array performs a resynchronization. During the resyn- chronization, host I/O is queued to the host, which can create a backlog of requests for processing when the array is able to respond. The resynchronization may take less than a minute, but the backlog of requests may take time to clear and some host I/O may experience timeouts. The HP recommended driver timeout values are 60 seconds. When an I/O timeout occurs, multipath drivers will typically retry the I/O request on the same path or an alternate path. There are operating system and HBA dependencies on these algorithms. The retry algorithms may extend the total amount of time that a host I/O will be allowed to complete successfully beyond the 60 second timeout values. Note that some applications may be more sensitive to I/O delays. • Processing host I/O during the upgrade-The upgrade competes with host requests that have a higher priority, which can affect the upgrade outcome. To ensure that host I/O from business critical applications or operations is minimized during the upgrade, select a time when a host I/O disruption would minimize the impact to the business operations. • Communicating with HP Command View EVA-The first steps of the upgrade have a lower priority than host I/O. The upgrade can extend beyond the timeout values expected by HP Command View EVA, which may cause loss of management connectivity to the EVA. A HP Command View EVA discover operation may be required to re-establish communication with the EVA. XCS 09534000 upgrade considerations Before upgrading to XCS 09534000, consider the following: • If the array contains solid state disk (SSD) drives, the upgrade must be done offline. • When upgrading from XCS 09522000 or earlier, an online controller software upgrade is not supported on an array connected to a host with applications using persistent reservations (for example, Windows Hyper-V with Cluster Shared Volumes, Windows 2008 SQL clusters, or Symantec Veritas Cluster Services with IO fencing). When upgrading from XCS 0953x000, an online upgrade is supported. • HP test results show that upgrading the controller software online on an EVA containing Vraid0 LUNs can marginally increase the possibility of Vraid0 LUN data unavailability during the upgrade. To reduce this increased risk, either halt all host I/O to Vraid0 LUNs or migrate the LUNs to another Vraid level before the upgrade. • The EVA resynchronization time is affected by allocated capacity plus the number of virtual disks and DR groups. Allocated capacity is the total of all virtual disks, snapshots, snapclones, mirrorclones, containers, and data replication log disks. When performing an online upgrade, to ensure a resynchronization time of less than 60 seconds, the total allocated capacity and number of virtual disks that are configured on an EVA should be one of the following: • less than 400 TB allocated capacity and less than 512 virtual disks created • less than 300 TB allocated capacity and less than 1024 virtual disks created • less than 125 TB allocated capacity and less than 2048 virtual disks created Use the following formula to calculate the total allocated capacity: HP StorageWorks Enterprise Virtual Array Updating Product Software Guide (XCS 09534000) 25

  • 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
  • 29
  • 30
  • 31
  • 32
  • 33
  • 34
  • 35
  • 36
  • 37
  • 38
  • 39
  • 40
  • 41
  • 42
  • 43
  • 44
  • 45
  • 46
  • 47
  • 48
  • 49
  • 50
  • 51
  • 52
  • 53
  • 54
  • 55
  • 56
  • 57
  • 58
  • 59
  • 60
  • 61
  • 62
  • 63
  • 64
  • 65
  • 66
  • 67
  • 68
  • 69
  • 70
  • 71
  • 72
  • 73
  • 74
  • 75
  • 76
  • 77
  • 78
  • 79
  • 80
  • 81
  • 82
  • 83
  • 84
  • 85
  • 86
  • 87
  • 88
  • 89
  • 90
Evaluating your configuration
The following factors can affect the success of an online upgrade. Consider these factors when
determining if an online upgrade is appropriate for your environment.
•
Queuing host I/O
—
After the upgrade, the array performs a resynchronization. During the resyn-
chronization, host I/O is queued to the host, which can create a backlog of requests for processing
when the array is able to respond. The resynchronization may take less than a minute, but the
backlog of requests may take time to clear and some host I/O may experience timeouts. The HP
recommended driver timeout values are 60 seconds. When an I/O timeout occurs, multipath
drivers will typically retry the I/O request on the same path or an alternate path. There are oper-
ating system and HBA dependencies on these algorithms. The retry algorithms may extend the
total amount of time that a host I/O will be allowed to complete successfully beyond the 60 second
timeout values. Note that some applications may be more sensitive to I/O delays.
•
Processing host I/O during the upgrade
—
The upgrade competes with host requests that have a
higher priority, which can affect the upgrade outcome. To ensure that host I/O from business
critical applications or operations is minimized during the upgrade, select a time when a host I/O
disruption would minimize the impact to the business operations.
•
Communicating with HP Command View EVA
—
The first steps of the upgrade have a lower priority
than host I/O. The upgrade can extend beyond the timeout values expected by HP Command
View EVA, which may cause loss of management connectivity to the EVA. A HP Command View
EVA discover operation may be required to re-establish communication with the EVA.
XCS 09534000 upgrade considerations
Before upgrading to XCS 09534000, consider the following:
•
If the array contains solid state disk (SSD) drives, the upgrade must be done offline.
•
When upgrading from XCS 09522000 or earlier, an online controller software upgrade is not
supported on an array connected to a host with applications using persistent reservations (for ex-
ample, Windows Hyper-V with Cluster Shared Volumes, Windows 2008 SQL clusters, or Symantec
Veritas Cluster Services with IO fencing). When upgrading from XCS 0953x000, an online upgrade
is supported.
•
HP test results show that upgrading the controller software online on an EVA containing Vraid0
LUNs can marginally increase the possibility of Vraid0 LUN data unavailability during the upgrade.
To reduce this increased risk, either halt all host I/O to Vraid0 LUNs or migrate the LUNs to an-
other Vraid level before the upgrade.
•
The EVA resynchronization time is affected by allocated capacity plus the number of virtual disks
and DR groups. Allocated capacity is the total of all virtual disks, snapshots, snapclones, mirror-
clones, containers, and data replication log disks. When performing an online upgrade, to ensure
a resynchronization time of less than 60 seconds, the total allocated capacity and number of vir-
tual disks that are configured on an EVA should be one of the following:
•
less than 400 TB allocated capacity and less than 512 virtual disks created
•
less than 300 TB allocated capacity and less than 1024 virtual disks created
•
less than 125 TB allocated capacity and less than 2048 virtual disks created
Use the following formula to calculate the total allocated capacity:
HP StorageWorks Enterprise Virtual Array Updating Product Software Guide (XCS 09534000)
25