Synergy Oneview tasks timeout

[CSKA]

Hello,

I'm not sure how to express my feedback regarding task management in OneView but see a lot of experts from HPE OneView department is contributing here, so get decision to open a new topic.

These days, I’ve received incidents regarding network ports down and start investigating the environment. As I’m remote administrator, I was not able to verify the equipment physically and my options were limited.

We have 2 IC modules VC SE 100Gb F32, installed in bays 3/6 and just the ports from bay6 were down. First of all, I’ve tried to refresh, then reapply the configuration and found that all tasks on that VC are running very slow. Seems like there isn’t proper timeout specified in the OneView logic.

For example, reapply configuration takes 8 hours to timeout and generate an error message. As VC was unresponsive, I’ve sent e-fuse command. Sub-task remove was immediately but adding was stuck again in hours. During that time, I was able to move all workloads on other frame, so finally I’ve got decision to delete the whole LE and to create a new one. So now, the delete process stuck again on subtask release the IC for more than 10h.
There is no way to cancel the task, no way to specify timeout before you submit the task. Meanwhile I found that seems like VC is broken as is not responding even to IP registration. Device can’t get link-local IPv6 address, that’s why OneView is not able to communicate with it. Just to be on a save side, both FLM are reset, communication to all other devices and linked enclosures is fine also. As frames are in secure location, I managed to get onsite contact with delay at the second day, now I have confirmation from the guy onsite that no any lights on the VC, seems like it is dead. I will open support case to HPE for part replacement, but I can’t stop any of the stuck tasks. Any advice on task management are welcome. To HPE colleagues, please think about it to get any instruments for the end users to stop stuck tasks or to implement proper timeout algorithms.

OneView Version - 8.10.00-0472141

BR,
Philip

[DanCernese]

Sorry for your experienece, it was a cascade of worsening conditions once the VC module went dead.  It is true that HPE OneView does not offer many situations where you can cancel operations.

As for the stuck tasks, HPE support has tools that, once granted access to the internals of the composer, can stop them.  Alternatively it seems that little harm would come of simply restarting.  If you restart and tasks remain, then only HPE support can clean those up manually.

You may consider two cases, one for part replacement and the other for composer cleanup if the restart doesn't do it.

[CSKA]

Thanks Dan,

I have support case logged, so currently is waiting support engineer to verify the logs and call me back to enter in MM console and delete the stucked task. On the other stream we will work for VC replacement.

BR,

Philip.