VxFlex OS ESXi SDSs disconnect on NVDIMM Ready nodes when adding disks

Summary: VxFlex OS ESXi SDSs disconnect on NVDIMM Ready nodes when adding disks

This article applies to This article does not apply to This article is not tied to any specific product. Not all product versions are identified in this article.

Symptoms

VxFlex OS ESXi SDSs on Ready nodes with NVDIMM disconnect when adding disks to the accelerated storage pool and may remain in a disconnected state. The SDS will begin to disconnect when a certain threshold of storage capacity is reached. This threshold is different depending on the SVM configuration.

The user may also see that disks enter an error state, the SDS may get stuck in a loop of disconnecting and reconnecting, or it may remain disconnected.

The user will also see alerts like this in the affected SDS's trc logs.

 trc.0

 17/12 14:44:07.491992 0x7f0f815addb8:NvDev_AllocNvBufByKeyInt:03305: Will allocated buffer by key:1104 of size:139856 on device id:7dff079e0001000a
17/12 14:44:07.492110 0x7f0f815addb8:NvDev_AllocNvBufByKeyInt:03368: Allocated buffer by key:1104 of size:139856 on device id:7dff079e0001000a
17/12 14:44:07.492112 0x7f0f815addb8:_AFS_NvramFileHardenDictEntry:05127: Wrote dict entry 36 for file (0x5 0x80b0005) on device 7dff079e0001000a nvBufId 1104 rc SUCCESS
17/12 14:44:07.492114 0x7f0f815addb8:_nvramBackend_FileCreate:05922: AFS File (0x5 0x80b0005) creation on NVRAM device 7dff079e0001000a done with rc = SUCCESS
17/12 14:44:07.492115 0x7f0f815addb8:AFS_FileCreate:03247: Creating AFS file (0x5 0x80b0005) ended with rc 65
17/12 14:44:07.492117 0x7f0f815addb8:_nvramBackend_DevInfo:05722: Retrieved device info on device 7dff079e0001000a . Maximum user space 30100 MB, free user space 26611 MB.
17/12 14:44:07.492147 0x7f0f815addb8:mosShmUmt_CreateOs:00152: shm_open failed, name emc_scaleio_spef_dev_9078983886427455496_comp_CHANGE_LOG_file_b_clbuckets, errno 17,size 23568
17/12 14:44:07.492149 0x7f0f815addb8:mosShmUmt_Create:00699: WARNING: SHM OS layer couldn't create shm file
17/12 14:44:07.492272 0x7f0f815addb8:mosDbg_SignalHandler:00712: ---Termination due to signal 7. PID 21557 Faulting address 0x7f0fc50c0000. errno 0---

17/12 14:44:07.492277 0x7f0f815addb8:contNet_AbnormalExitCK:02433: Will pause network
17/12 14:44:07.492279 0x7f0f815addb8:net_Pause:02014: Net paused 1 (reversible 0)

Cause

This issue happens because of not provisioning enough RAM to the SVM that the SDS resides on.

Resolution

To resolve this issue, change the amount of RAM allocated to the SVM.
To calculate the amount of RAM required on a per node basis use the following formula.

Minimum_RAM_capacity_in_GB  =  10 + ((100 * Number_of_accelerated_drives_on_node) + (550 * Total_node_accelerated_capacity)) / 1024

Alternatively, User can follow this table (Values below are for the entire accelerated SP)

 
FG capacity Required
NVDIMM
capacity
Required RAM
capacity
(rounded)
Additional
memory for
MDM, LIA, and
SVM OS
Total RAM
(rounded)
51.2 TB 32 GB NVDIMM
(in SVM it is 31 GB)
X > 41 GB MDM:5.4GB
LIA: 350 MB
OS Base: 1 GB
Buffer: 1 GB
53 GB
51.2 TB < X < 96 TB 64 GB NVDIMM
(in SVM it will be 62 GB)
41 GB < X < 64.5 GB 53 GB < X < 73 GB
96 TB < X < 128 TB
(122.88 is the actual limit)
96 GB NVDIMM
(in SVM it will be 93 GB)
64.5 GB < X < 81.5 GB 87 GB

Additional Information

Affected Products

VxFlex Ready Nodes, VxFlex Product Family, VxFlex Ready Node, VxFlex Ready Node R640, VxFlex Ready Node R740xd, VxFlex Ready Node R840
Article Properties
Article Number: 000168850
Article Type: Solution
Last Modified: 25 Nov 2024
Version:  3
Find answers to your questions from other Dell users
Support Services
Check if your device is covered by Support Services.