qla2xx failure

Latest response

qla2xxx_18_dpc: page allocation failure: order:4, mode:0x80d0

Responses

Hi Mike,

We are missing lot of basic information to help you.

Firstly, this Red Hat document nicely describes common reasons for page allocation failures:

https://access.redhat.com/articles/1360023

a) What is the version of the operating system and kernel?

b) Did the error just started happening, or is it an ongoing problem?

c) What is the result of the error (file system cannot be mounted, write errors on the disk, application crashing, or something else)?

Best wishes,

Dusan Baljevic (amateur radio VK2COT)

7.7 3.10.0-1062.1.2.el7.x86_64 This is the first time we have seen this happen. The application went down (oracle) because the SAN shares were unavailable

Hi,

So it is RHEL 7.7.

Were there any other errors at the time of problem? I would expect more errors to show in logs.

As well, do you have any performance stats around the time when problem occurred (like memory fully used and swapping started)?

What is the file system type where Oracle app resides?

Out of interest, what do these commands show for your disk where Oracle app resides now (when the system is fully operational):

# sg_persist -d <diskdev> -c

# lsscsi -i -s -w -x

# ls -v -1c /dev/sd*[!0-9] | xargs -I {} sh -c 'echo -n "{} : " ; /lib/udev/scsi_id --whitelisted --device={}'

It would be good to open case with Red Hat...

Regards,

Dusan Baljevic (amateur radio VK2COT)

You also might consider enabling qla2xx debugging in non-prod environment, as per Red Hat knowledge article

https://access.redhat.com/articles/337813

Dusan Baljevic (amateur radio VK2COT)