SCv3000 health OK, but fault indicator still yellow flashing

Hello polezhaevdmi,

If you select the alerts tab what alerts or errors are listed?  What is your current SCOS version?

Hello, Sam!

System looks completely free of alerts now: even there are no connectivity alerts regarding switched off servers (gone during 2 last days).

Regarding the SCOS version. System reports "7.4.21.4", but I remember I updated it from factory "7.4.21.4" to something like "Update 1" or "7.4.21.4.1" during deployment session about 1 month ago (as the update appeared available). Possibly, the version report displays 4 numbers only.

PS> And yes, the fault indication - is still active, we checked this today again.

Hello polezhaevdmi,

I would see if you can get a time to reboot your system & see if the error light comes back & it reports an error in the alerts tab.

Thank You for advice, @DELL-Sam L !

OK, it will require about up to 3-4 weeks to negotiate and perform the reboot, I will write the result. 

And, to be accurate: Sam, please, confirm You're talking about complete procedure with storage "(stop production, indeed) -> shut down storage -> switch off PSU power -> remove AC cables -> wait for 2+ minutes -> re-seat AC cables back -> power on storage -> (start production, indeed)" (like here). Not the "soft reboot".

Cheers!

Hello polezhaevdmi,

I would try & do the soft reboot first to see if it resolves your issue.  If it doesn’t resolve the error light, then you are going to do the shutdown and start process.

Hello, Sam!

A lot of thanks, the Plan is clear (however, not simple: as to see the front panel one must go for ~500 meters and extra security point ), I will write the result in a couple of weeks.

Cheers!

Hello again!

Here are the results:

- The 'soft reset' attempt vas in vain;

- The 'complete hardware power cycle' was in vain too. Still the fault indicator flashes while web-console status is OK and no new alarms. I attached the summary screenshot again. In addition, we created a video-log of the system start process. The alarm indicator starts to be active at 04:40 time point approximately.

Hello!

Yes, Sam, the support for that system is active, I will continue with SR, and drop some information regarding result here. Great thanks for Your efforts!

While waiting for SupportAssist files from first Customer, I eventially found another system with same effect. Different Company, different storage (SC5020F), different firmware - but effect looks the same. The common things among them I'm able to notice:

- Both organizations have server cluster objects to provide the access to shared storage;

- Both organizations had some servers connectivity lost, so, the cluster objects reported an alarm, which were propagated to storage level;

- Both organizations solved the lost server warning (by reviving the server or removing unneccessary server object);

- It appeared me, who visited both organizations and was asked 'why THAT blinks yellow?'.

One SR is initiated, other one will be soon.

OK, LED issues at both systems were fixed with Compellent Support Team without downtime. By the way, the issues appeared not simple: Level-2 Engineer was invited for consultation. The second system (SC5020F) appeared to require additional intervension (in deduplication 'mechanics'), we are still in resolutuion process. Still, the LEDs are fine now. Thus, if anyone run into such effect again: do not hesitate and open Service Request. And - thanks to Community, which helped to prepare all data for Support Collegues!

I am facing the same problem with the Dell SCv3020. The health light is indicating amber, but there is no alarm on the Dell Storage Manager.

Hi,

Was there any updates within the storage system? If you have already checked the storage manager there is no alert on error or issues then try running the console command# fm pfm showledfault on both controllers. If the output shows status Clear, then it might be from one of the controllers. If that is the issue, you may need to contact support to raise a case for a remote connection. As the post owner mentioned, the support need to remote into the controller to investigate.