Server with 3PARdata remote copy LUNs shows huge number of multipath path failure errors and fails to boot
Issue
-
Server with 100+ 3PARdata remote copy LUNs shows huge number of multipath path failure errors and fails to boot:
Aug 22 11:43:22 testhost multipathd: sdb: mark as failed Aug 22 11:43:22 testhost multipathd: mpathap: remaining active paths: 1 Aug 22 11:43:22 testhost multipathd: sdk: mark as failed Aug 22 11:43:22 testhost multipathd: mpathap: Entering recovery mode: max_retries=18 Aug 22 11:43:22 testhost multipathd: mpathap: remaining active paths: 0 Aug 22 11:43:22 testhost multipathd: sdn: mark as failed Aug 22 11:43:22 testhost multipathd: mpathag: Entering recovery mode: max_retries=18 Aug 22 11:43:22 testhost multipathd: mpathag: remaining active paths: 0 Aug 22 11:43:22 testhost multipathd: sdf: mark as failed Aug 22 11:43:22 testhost multipathd: mpathbf: Entering recovery mode: max_retries=18 Aug 22 11:43:22 testhost multipathd: mpathbf: remaining active paths: 0 Aug 22 11:43:22 testhost multipathd: mpathap: Entering recovery mode: max_retries=18 [...]Boot process gets hung after the above path failure errors.
-
Above errors and hang in boot process are observed only when the server is having active paths assigned from 3PAR storage controller in remote site and standby paths from local site.
-
The "active ghost" paths are discovered before "active ready" paths resulting in io errors and path failures when attempted use of the ghost
paths (alua Standby) is tried.
Environment
- Red Hat Enterprise Linux 7
- device-mapper-multipath
- 3PARdata LUNs configured with peer-persistence and remote copy
(Active paths assigned from 3PAR storage controller in remote site and standby paths from local site)
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.
