PIM - error SEOS_load: SEOS_syscall WASN'T loaded

[Atifah Abdul Latif]

Hi Support,

Below error prompted when we try to start PIM by using seload command

SEOS_load: SEOS_syscall WASN'T loaded

Below are the details of customer environment

PIM: 12.81.2919
OS: RHEL 7.4
Kernel: 3.10.0-693.el7.x86_64

Appreciate any advice. Thank you.

Regards,
Atifah

[Atifah Abdul Latif]

Hi Experts,

Appreciate any advice on how to move forward with this error. We have several servers with the same OS and PIM version, no issue to boot.

Thank you,
Atifah

[Joseph Lutz]

Atifah

Seos uses kernel modules. This means the module you try to load should support this kernel version. You are not using a module that supports this kernel version so you do need to upgrade. Before applying any kernel updates on your Linux installs you should validate the module supports that kernel release to avoid an outage or worse. Please update based on the latest release available from the webpage below (all higher builds include support from the lower builds)

Joe 

https://support.broadcom.com/web/ecx/support-content-notification/-/external/content/release-announcements/CA-PRIVILEGED-IDENTITY-MANAGER-ENDPOINT-COMPATIBILITY-MATRIX/16769

[Atifah Abdul Latif]

Hi @Joseph Lutz,

Thanks for your advice. However, there​ is another server with the same OS and kernel are booting normally with version 12.81.2919.
Only this one server has the issue. Is there any justification I can use to advise the customer to upgrade? because as of now other server run without issue.

Thank you,
Atifah

[Joseph Fry]

I would show the customer the compatibility matrix that Joe Lutz provided.  Just because it works on the other server does not mean it is supported.  In a production environment, you want to be running a supported configuration; not one that happens to work but is clearly unsupported.

[Joseph Lutz]

Atifah

I would tend to be a little more strict in the concepts of support. The fact that it loaded is not proof that it is working. One of the reasons why you could see the error you described is that someone attempted to restart the software but the module could not be unloaded cleanly and now is partially hung in the kernel. Rebooting the server may be enough to get it working at least to the level it was before.

If you had a sever crash because of this module did something wrong and opened a ticket with support we would simply tell you that it was too old to run on that kernel release and simply ask you to upgrade to the latest without any further review. If you and your client are simply ok with that you can continue to try to get it loaded or you could take the simpler approach and upgrade the software.

As Joe Fry mentioned, you do need to inform the client of this and provide them with the Matrix because that build of the endpoint is not supported at all, even for simple questions to support.

Joe Lutz

[Atifah Abdul Latif]

Thank you guys for the advice. Appreciate it.