VPLEX:由於儲存陣列的 scsi 檢查條件 B/44/00,儲存磁碟區處於嚴重故障狀態
Summary: 由於來自底層儲存陣列的 scsi 感知代碼 B/44/00,VPLEX 會將磁碟標記為失效。
Symptoms
VPLEX 儲存磁碟區處於嚴重故障狀態,而基礎磁碟區在陣列端狀況良好。
CLI 命令的範例輸出: ll /clusters/cluster-1/storage-elements/storage-volumes/<storage-volume name>
VPlexcli:/> ll /clusters/cluster-1/storage-elements/storage-volumes/Symm1234_0000 /clusters/cluster-1/storage-elements/storage-volumes/Symm1234_0000: Name Value ----------------------------- ------------------------------------------------ application-consistent false block-count 65798400 block-size 4K capacity 251G description - free-chunks [] health-indications [hardware dead] health-state critical-failure <<-- io-status dead itls 0x50001442607de710/0x5000097378028844/80, 0x50001442607de710/0x50000973780288c4/80, 0x50001442607de711/0x5000097378028844/80, 0x50001442607de711/0x50000973780288c4/80, 0x50001442707de710/0x5000097378028844/80, 0x50001442707de710/0x50000973780288c4/80, 0x50001442707de711/0x5000097378028844/80, 0x50001442707de711/0x50000973780288c4/80, 0x50001442a0100410/0x5000097378028844/80, 0x50001442a0100410/0x50000973780288c4/80, ... (16 total) largest-free-chunk 0B locality - operational-status error provision-type legacy storage-array-name EMC-SYMMETRIX-XXXXXXXXX storage-volumetype normal system-id VPD83T3:60000097000000000000000000000000 thin-capable false thin-rebuild true total-free-space 0B underlying-storage-block-size 512 use unusable used-by [extent_Symm1234_0000_1] vendor-specific-name EMC
儲存磁碟區處於 嚴重故障 狀態
VPlexcli:/> storage-volume summary SUMMARY (cluster-1) StorageVolume Name IO Status Operational Status Health State ---------------------------------------- --------- ------------------ ---------------- Symm1234_0000 dead error critical-failure <<--
Dial Home 的摘錄示例,突出顯示了死亡的邏輯單元:
<EventData><![CDATA[disk VPD83T3:60000970000000000000000000000000: disk died, no mirror exists
如果此儲存磁碟區不是分散式裝置的一部分,則主機將無法存取此虛擬磁碟區,從而導致資料不可用性的情況。
VPlex 韌體記錄顯示串流 SCSI/27 事件與 ASC 0x44 ascq 0x0 (B/44/00) 0xb asc 傳回 內部目標故障。
來自 firmware.log 的 SCSI/27 例項範例
128.221.253.37/cpu0/log:5988:W/"0060166fcbfe192045-1":3298782:<6>2016/11/11 06:50:57.94: scsi/27 tgt VPD83T3:60000097000000000000000000000000 cmd 0x2a status 0x2 valid 0 resp 0x70 seg 0x0 bits 0x0 key 0xb info 0x0 alen 10 csi 0x0 asc 0x44 ascq 0x0 fru 0x0 sks 0x0 128.221.253.37/cpu0/log:5988:W/"0060166fcbfe192045-1":3298783:<6>2016/11/11 06:50:58.00: scsi/27 tgt VPD83T3:60000097000000000000000000000000 cmd 0x2a status 0x2 valid 0 resp 0x70 seg 0x0 bits 0x0 key 0xb info 0x0 alen 10 csi 0x0 asc 0x44 ascq 0x0 fru 0x0 sks 0x0 128.221.253.37/cpu0/log:5988:W/"0060166fcbfe192045-1":3298784:<6>2016/11/11 06:50:58.16: scsi/27 tgt VPD83T3:60000097000000000000000000000000 cmd 0x2a status 0x2 valid 0 resp 0x70 seg 0x0 bits 0x0 key 0xb info 0x0 alen 10 csi 0x0 asc 0x44 ascq 0x0 fru 0x0 sks 0x0 128.221.253.37/cpu0/log:5988:W/"0060166fcbfe192045-1":3298785:<6>2016/11/11 06:50:58.30: scsi/27 tgt VPD83T3:60000097000000000000000000000000 cmd 0x2a status 0x2 valid 0 resp 0x70 seg 0x0 bits 0x0 key 0xb info 0x0 alen 10 csi 0x0 asc 0x44 ascq 0x0 fru 0x0 sks 0x0
執行 21 次重試,儲存磁碟區因為連續寫入失敗而被標示為失效。
128.221.253.37/cpu0/log:5988:W/"0060166fcbfe192045-1":3298797:<4>2016/11/11 06:51:01.42: amf/45 disk VPD83T3:60000097000000000000000000000000: write failure: marking this in-use disk dead 128.221.253.37/cpu0/log:5988:W/"0060166fcbfe192045-1":3298798:<2>2016/11/11 06:51:01.42: amf/97 disk VPD83T3:60000097000000000000000000000000: disk died, no mirror exists
Cause
儲存陣列針對寫入 I/O 要求 (0x2a傳回 B/44/00,陣列無法處理 I/O 要求,並以檢查條件 B/44/00 (內部目標故障) 回應
如果基礎儲存磁碟區是 VMAX 磁碟區,則可能會發生 I/O 故障,原因如下:
VMAX 中止 IO 的原因並未提供給 VPlex。發生內部目標故障的原因有很多,而且可能由非常廣泛的來源引起。
Resolution
傳送 scsi 檢查條件 B/44/00 至 VPLEX 的儲存陣列必須由個別陣列廠商進行調查。此問題是由於儲存陣列上的「內部目標故障」問題,而無法回應寫入 I/O 要求所觸發。
如果基礎陣列是 VMAX 陣列,則陣列可以傳送 scsi 檢查條件「B/44/00」,原因有多種。
可在 VPLEX 管理伺服器上執行下列 cli 命令,以取得受 B/44/00 檢查條件影響的邏輯單元清單:
grep "scsi/27" firmware.log* | grep 'key 0xb' |awk '{print $3, $5, $18, $19, $26,$27,$28,$29}'| sort | uniq -c | sort -nr
範例:
service@ManagementServer:~> grep "scsi/27" firmware.log* | grep 'key 0xb' | awk '{print $3, $5, $18, $19, $26,$27,$28,$29}'| sort | uniq -c | sort -nr
103474 scsi/27 VPD83T3:60000097000000000000000000000000 key 0xb asc 0x44 ascq 0x0
45254 scsi/27 VPD83T3:60000097000000000000000000000001 key 0xb asc 0x44 ascq 0x0
32908 scsi/27 VPD83T3:60000097000000000000000000000002 key 0xb asc 0x44 ascq 0x0
25581 scsi/27 VPD83T3:60000097000000000000000000000003 key 0xb asc 0x44 ascq 0x0
21043 scsi/27 VPD83T3:60000097000000000000000000000004 key 0xb asc 0x44 ascq 0x0
10809 scsi/27 VPD83T3:60000097000000000000000000000005 key 0xb asc 0x44 ascq 0x0
10651 scsi/27 VPD83T3:60000097000000000000000000000006 key 0xb asc 0x44 ascq 0x0
如果這是非 EMC 陣列,請與個別陣列廠商聯絡,以解決儲存陣列上存在的問題。
Additional Information
這不是 VPlex 的問題,這可能是後端陣列或光纖問題。
必須與儲存陣列團隊、連線能力或兩者接洽。