ESXI 6.0 VSAN 添加 第三組 DISK Group 失敗，造成IOPS 異常

VSAN 添加 第三組 DISK Group 失敗，造成IOPS 異常

因為會一直嘗試 Mount 異常的 HD，造成整個 VSAN crash

---

添加 DISK Group 失敗

發現慢到連 VC 都無法登入

上面的 VM GUEST 都慢到掛了

---

登入 ESXI ，確認 HD的 狀態，可以看到有五顆 HD 異常

[root@VM:~] localcli vsan storage list |grep CMMDS

   In CMMDS: true

   In CMMDS: true

   In CMMDS: true

   In CMMDS: true

   In CMMDS: true

   In CMMDS: true

   In CMMDS: true

   In CMMDS: true

   In CMMDS: true

   In CMMDS: true

   In CMMDS: true

   In CMMDS: false

   In CMMDS: false

   In CMMDS: true

   In CMMDS: true

   In CMMDS: true

   In CMMDS: false

   In CMMDS: false

   In CMMDS: true

   In CMMDS: true

   In CMMDS: true

   In CMMDS: true

   In CMMDS: false

   In CMMDS: true

照群組，列出有異常的 HD，

[root@VM:~] localcli vsan storage list |grep "CMMDS: false" -B6 |grep naa

   Display Name: naa.50000398082a98e5

   VSAN Disk Group Name: naa.5002538a07159ae0

   Display Name: naa.50000398082a92b9

   VSAN Disk Group Name: naa.5002538a07159ae0

   Display Name: naa.50000398082ab1e9

   VSAN Disk Group Name: naa.5002538a07159ae0

   Display Name: naa.50000398082a7495

   VSAN Disk Group Name: naa.5002538a07159ae0

   Display Name: naa.50000398082ab169

   VSAN Disk Group Name: naa.5002538a07159ae0

將這個異常的 Disk Group Unmount

[root@VM:~] esxcli vsan storage diskgroup unmount -s naa.5002538a07159ae0

unmount 後，可以發現這些 HD 都顯示異常，再將顯示為 unmont 的 Disk Group 刪除

---

建議：

HD 若不確定裡面是否有資料，建議要做 Erase Partitions ，將HD 清空

可以點 Turn Locator LED on 來判斷 實體 HD，避免出錯

---

原因 

https://kb.vmware.com/s/article/2146495

The issue of disk group creation failure on a vSAN ESXi host occurs under these conditions:

• There are 3 or more disk groups per ESXi host.
• There are large cache tier drives with a capacity of approximately 1TB or greater.

• 剛好要在ESXI 上添加 第三組 Disk Group
• SSD 1.9 TB，大於 1TB

---

解決方案 -- 

---

This issue is resolved in these releases:

• ESXi 6.0 Update 3, available from VMware Downloads
• ESXi 6.5 GA, available from VMware Downloads

To work around this issue, increase the heap memory size on the ESXi host with these steps:

1. Connect to the affected ESXi host with an SSH session.
2. Run this command:
   esxcfg-advcfg -s 1024 /LSOM/heapSize
   Note: This will change the default heap memory size from 256 to 1024
3. Reboot the ESXi host.
4. Attempt the Disk group creation again.

下指令前，先確認目前參數設定

esxcli system settings advanced list -o "/LSOM/heapSize"

下指令後，再確認目前參數設定 ，已變更成 1024

esxcfg-advcfg -s 1024 /LSOM/heapSize

---

Disk Group 添加成功，每顆 HD 看來都正常

[root@VM:/var/log] esxcli vsan storage list |grep CMMDS

   In CMMDS: true

   In CMMDS: true

   In CMMDS: true

   In CMMDS: true

   In CMMDS: true

   In CMMDS: true

   In CMMDS: true

   In CMMDS: true

   In CMMDS: true

   In CMMDS: true

   In CMMDS: true

   In CMMDS: true

   In CMMDS: true

   In CMMDS: true

   In CMMDS: true

   In CMMDS: true

   In CMMDS: true

   In CMMDS: true

   In CMMDS: true

   In CMMDS: true

   In CMMDS: true

   In CMMDS: true

   In CMMDS: true

   In CMMDS: true