OVM : iSCSI LUN on Guest VM has failed paths causing RAC node reboot (Doc ID 2372276.1)

Last updated on AUGUST 23, 2022

Applies to:

Symptoms

On DomU, GuestOS

• LUN has failed paths on Guest and after few weeks the gust reboots
• The Guest is a RAC node and have LUNs coming via iSCSI

# multipath -ll
sm_OraRacPrdOmnilabRedoTmp1_asm1 (360a9800044344243643f4547476a4d37) dm-126 NETAPP,LUN
size=25G features='4 queue_if_no_path pg_init_retries 50 retain_attached_hw_handle' hwhandler='0' wp=rw
`-+- policy='round-robin 0' prio=2 status=active
  |- 3:0:0:62 sdec 128:64 active ready running
  `- 2:0:0:62 sddo 71:96 failed faulty running <<<------ Failed path
sm_OraRacPrdWinrestRedoTmp1_asm1 (360a9800044344243643f4547476a4872) dm-58 NETAPP,LUN
size=10G features='4 queue_if_no_path pg_init_retries 50 retain_attached_hw_handle' hwhandler='0' wp=rw
`-+- policy='round-robin 0' prio=2 status=active
  |- 3:0:0:40 sdcj 69:112 active ready running
  `- 2:0:0:40 sdbt 68:112 failed faulty running <<<------ Failed path
...

• Ping to the iSCSI target IP with failed paths fails
• Ping to working iSCSI target fails but  iSCSI port is reachable ; that is why one path is working

# nc -zv xx.xx.xx.xx 3260
Connection to xx.xx.xx.xx 3260 port [tcp/iscsi-target] succeeded!

• tcpdump on VIF interface not showing ICMP reply , but ICMP reply reaches up to bridge

Changes

 No change

Cause

My Oracle Support provides customers with access to over a million knowledge articles and a vibrant support community of peers and Oracle experts.