7
Solution
To resolve the issue:
1. Do not disable all internal model controls (IMCs) for CPU 1 or CPU 2 if NUMA is enabled, or
disable NUMA if you want to disable all IMCs for CPU 1 or CPU 2.
2. If the issue persists, contact H3C Support.
An error occurred on an NVMe drive and the drive
went offline after a managed hot plug
Symptom
If an OS earlier than Linux Kernel 4.3 is used, an error occurred on an NVMe drive and the drive went
offline after a managed hot plug.
Solution
This issue is caused by payload inconsistency between the NVMe drive and PCIe bridge device after
a managed hot plug of the NVMe drive.
To resolve the issue:
1. Use Linux Kernel 4.3 because a code for negotiating the payload has been added to this
version.
2. To continue using an OS earlier than Linux Kernel 4.3, use the following methods:
Method 1:
a. Use the setpci command to set the value of max_payload_size in the NVMe register to be
the same as that of the bridge device.
b. Display the value of payload on the PCIe bridge device. You must set the value of
maxpayload in DevCtl of the NVMe drive to be the same as the value of this payload.
Figure 243 Displaying the value of the payload on the PCIe device
c. Display the value of the DevCtl field in the register, and then set the value of the
maxpayload field. As shown in the following figure, if the value of the DevCtl field is
00002810, set the value of the maxpayload field to 128B. To set it to 256B, the value of the
DevCtl field must be 0x2830.
To Next Page
Loading...
Need help?
General
