PowerFlex: Disaccoppiamento continuo SDS MTU non corrispondente

Un aggiornamento a vCenter non è riuscito ed è stata implementata una nuova istanza di vCenter. Le SVM esistenti sono state aggiunte al nuovo vCenter e quindi al DVS (distributed virtual switch) appena creato. Una volta completata questa operazione, un SDS ha continuato a entrare in uno stato disaccoppiato/disconnesso L SDS viene visualizzato come disconnected, Join-Pending in scli --query_all_sds prodotto:

SDS ID: b7g3bd870000000aName: SIO-ESX1 State: Disconnected, Join-Pending IP: 172.21.55.110 Port: 7072 Version: 2.0.13000

Il registro eventi MDM mostra che l SDS si disaccoppia continuamente, entra in modalità di raffreddamento e quindi si riconnette.

2017-10-09 12:35:53.155 SDS_DECOUPLED ERROR SDS: SIO-ESX1 (id b7g3bd870000000a) decoupled. 2017-10-09 12:35:53.156 SDS_IN_COOL_DOWN WARNING 
SDS: SIO-ESX1 (ID b7g3bd870000000a) will disconnect from MDM for 15 seconds failed to reconnect multiple times 2017-10-09 12:36:08.155 
SDS_RECONNECTED INFO SDS: SIO-ESX1 (ID b7g3bd870000000a) reconnected 2017-10-09 12:36:13.161 SDS_RECONNECTED INFO SDS: SIO-ESX1 
(ID b7g3bd870000000a) reconnected 2017-10-09 12:36:18.255 SDS_RECONNECTED INFO SDS: SIO-ESX1 (ID b7g3bd870000000a) reconnected 
2017-10-09 12:36:23.259 SDS_RECONNECTED INFO SDS: SIO-ESX1 (ID b7g3bd870000000a) reconnected 2017-10-09 12:36:28.255 SDS_RECONNECTED 
INFO SDS: SIO-ESX1 (ID b7g3bd870000000a) reconnected ... 2017-10-09 12:38:41.153 SDS_DECOUPLED ERROR SDS: SIO-ESX1 (id b7g3bd870000000a) 
decoupled. 2017-10-09 12:38:41.155 SDS_IN_COOL_DOWN WARNING SDS: SIO-ESX1 (ID b7g3bd870000000a) will disconnect from MDM for 15 seconds
 failed to reconnect multiple times 2017-10-09 12:38:56.159 SDS_RECONNECTED INFO SDS: SIO-ESX1 (ID b7g3bd870000000a) reconnected
 2017-10-09 12:39:01.164 SDS_RECONNECTED INFO SDS: SIO-ESX1 (ID b7g3bd870000000a) reconnected 2017-10-09 12:39:06.260 SDS_RECONNECTED 
INFO SDS: SIO-ESX1 (ID b7g3bd870000000a) reconnected 2017-10-09 12:39:11.265 SDS_RECONNECTED INFO SDS: SIO-ESX1 (ID b7g3bd870000000a) 
reconnected

I registri di traccia MDM mostrano che l SDS ha problemi di comunicazione. L SDS si connette e poi si disconnette:

09/10 12:39:11.261955 0x7f66b2a32eb0:tgtMgr_CheckMsgReturnedRC:01728: Message RC is TIMEOUT, RC is COMMUNICATION_ERROR 09/10 12:39:11.261972
 0x7f66b2a32eb0:tgtMgr_ConfigureTgt:05190: TGT_CONFIG Failed to update new TGT b9f3bd870000000a on all other TGTs RC: COMMUNICATION_ERROR
 09/10 12:39:11.261981 0x7f66b2a32eb0:tgtMgr_ProcessComplete:07188: Tgt: b7g3bd870000000a RC=COMMUNICATION_ERROR type: UP upDownState: 
DOWN processState: UP_INPROGRESS mdmTgtConGen: 51897 upPend: 1 downPend: 0 devErrPend: 0 wasDown: 1 createIP: 0 reconstructIP: 0 09/10
 12:39:11.261994 0x7f66b2a32eb0:keepalive_SetTgtFenced:00331: TGT: b7g3bd870000000a 09/10 12:39:11.262015 0x7f66b0902eb0:repExtent_IO:02976:
 Writing the repository to disk and cluster. Extent: 0 Page: 186 Offset: 761856 Size: 65536 09/10 12:39:11.263420 
0x7f66b2a32eb0:tgtMgr_InitiateProcessUnlocked:07027: Tgt: b7g3bd870000000a processType: UP upDownState: DOWN processState: IDLE 
mdmTgtConGen: 51897 upPending: 0 downPending: 0 wasDown: 1 createIP: 0 reconstructIP: 0 aboutToBeRemoved: 0 09/10 12:39:11.263472
 0x7f66b2a71eb0:tgtMgr_HandleWorkReq:04511: TgtId: b7g3bd870000000a starting Up process 09/10 12:39:11.263488 0x7f66b0926eb0:
repExtent_IO:02976: Writing the repository to disk and cluster. Extent: 0 Page: 186 Offset: 761856 Size: 65536 09/10 12:39:11.264680 
0x7f66b2a71eb0:keepalive_SetTgtListenOnly:00377: TGT: b7g3bd870000000a 09/10 12:39:11.264689 0x7f66b2a71eb0:mdmTgtMsg_SendSyncAddMdm:00505
: TGT_CONFIG TgtId: b7g3bd870000000a MdmId: 47e9330a1b66a52e mdmTgtConnectionGenNum: 51898 bForceClean: 0 bQuiesce: 0 quiesceGen: 1 
09/10 12:39:11.265081 0x7f66b2a71eb0:mosEventLog_PostInternal:00590: New event added. Message: "SDS: SIO-ESX (ID b7g3bd870000000a) 
reconnected". Additional info: "" Severity: Info 09/10 12:39:11.265091 0x7f66b2a71eb0:mdmTgtMsg_SendReconfStart:00571: TGT_CONFIG 
TgtId: b7g3bd870000000a

Non si tratta di un problema di ScaleIO, ma di un problema di configurazione della rete ESXi.

Modificare l'uplink DVS (Distributed Virtual Switch) e impostare la dimensione MTU su 9000 per coerenza con tutte le impostazioni MTU di rete.
Verificare che tutti i componenti nel percorso di rete siano impostati su 9000 MTU.