Rebuild VPLEX Distributed Volumes

0
188
REBUILD
A REBUILD WILL OCCUR WHEN
=========================
*Logging failed, mirror will be marked out of date.
*whenever a new mirror is added to device, as in the case of a migration.
*whenever an existing device is marked out of date. 
*A distributed raid-1 has one leg marked out of date due to storage issues or inter cluster
link issues.
*A logging rebuild will occur when a mirror (leg) of a distributed-device is removed from the
source
VPlexcli:/> rebuild show-transfer-size *     This command will return transfer rate for all
devices
VPlexcli:/> rebuild show-transfer-size --devices *     This command will return device name
and transfer size
device name  transfer-size
-----------  -------------
ddev_1       2M
ddev_10      2M
ddev_11      2M
ddev_12      2M
ddev_13      2M
ddev_7       2M
ddev_8       2M
ddev_9       2M
Default       128k    This has been tested and works best
Max size    128M
Set the transfer-size on a specified device to 1M:
======================================
VPlexcli:/> rebuild set-transfer-size --devices /clusters/cluster-1/devices/testdevice --limit
1M
** alternatively:  VPlexcli:/distributed-storage/distributed-devices/DD_1234>set transfer-size
1M
setting transfer size at component level
===============================
VPlexcli:/distributed-storage/distributed-devices/DD_1234/distributed-device-components/device
_Symm1234_1234_1> set transfer-size 1M
Set the transfer-size for all devices to 2M:
VPlexcli:/> rebuild set-transfer-size /clusters/*/devices/* 2M
Set the transfer-size for all distributed devices to 10K:
VPlexcli:/distributed-storage/distributed-devices> rebuild set-transfer-size * 10k
Be sure to note if 'throughput' increases after increasing the transfer size.
It's not always the case that increasing the transfer rate will speed up the rebuilds.

LEAVE A REPLY

Please enter your comment!
Please enter your name here