Hardware

Hi

Check by reseting the controller

Sun Storage 2500, 2500-M2, and 6000 Arrays: Performance is Slow on an Apparently Optimal Array (Doc ID 1674848.1)

And other advise is,

If it is SMART then you can reset the battery age without replacement and you can make active

Battery Expiration is no longer cause for replacement on Storage 25x0/6x80 Arrays. (Doc ID 1171913.1)

How to reset battery Age:

How to Reset the Cache Backup Battery Age for Sun Storage 2500, 2500-M2, and 6000 Arrays (Doc ID 1021695.1)

Using the CAM GUI:

    1.  Open a browser the the CAM host:  https://cam_server:6789

    2.  Expand "Storage Systems" Menu

    3.  Expand your array name menu

    4.  Click on the Service Advisor link at the top.

    5.  Expand the Backup Battery Pack Removal/Replacement menu.

    6.  Click on the Reset the battery age link.

    7.  Click the Reset button.

Incase non-smart batteries then you need to replace the failed battery for optimal results.

Thanks

Vedha

Dear Vedha,

thanks for your reply. We tried to reset the Battery Age but it does not work since in our case the State is Enabled but the Status is Error

In my understanding the cache should be able to work even without batteries if we set the option "Write Cache without Batteries Enabled" to true. Right? In our case it does not. It still says "Write Cache Active: False". We tried restarting the storage but it still does not work.

We will order new batteries but we have to find a fix for the poor performance until we receive the batteries. Any help would be greatly appreciated!

Thanks

Hi,

have you done the changes for write cache via CAM BUI? Normally nothing more is to do here

Perhaps you want to try it to change via commandline as follows and please be aware that you have to do this settings for each Volume:

Using the CAM GUI

1. Open a browser to the CAM host: https://<server IP>:6789.
2. Expand "Storage Systems" menu.
3. Expand your array name menu.
4. Click the "Volumes" menu.
5. Click your volume name.
6. Change the "Write Cache Without Batteries Enabled" from FALSE to TRUE.

Using the CAM command line

1. Set the "Write Cache Without Batteries" to TRUE:
   
   "sscs modify -a <arrayname> -b true volume <volume name>"

Where the location for sscs is:

Solaris: /opt/SUNWstkcam/bin/
Linux: /opt/sun/cam/bin/
Windows: C:\Program Files\Sun\Common Array Manager\bin

best regards

Uwe

NOTE: additional I want to make you aware of the risk of loosing data in cache when write cache without batteries is enabled during a power outage

Hi,

"Write Cache without Batteries Enabled"

Yes,This will enable the cache back but there are risks to lose the data in cache in case of a power outage.

Further investigation is needed, Please raise a SR with Support Team.

Thanks

Vedha

Dear Vedha,

  we have already make the change  and pressed save on all volumes. It says Change has been successful but the Write Cache Active is still false. I don't understand why is that? Shouldn't I get the Write Cache Active back to true?

Thanks,

Savvas

Dear Uwe,

we did the changes via the Web GUI but it still not working. Any ideas why this is happening?

Thanks

Hello,

Since both batteries are offline, I'm suspecting that you have a different issue with this array and you will likely not resolve it by forcing the cache manually. At this stage I think it will be required to open a Service Request with the Oracle support so that we can analyze in depth the diagnostic data from your array.

Regards

Nicolas

Dear Nicolas,

  thanks for your reply. We will do so.

Regards,

Savvas

Dear Nicolas,

  unfortunately I found that since the storage is too old we don't have a maintenance contract active with Oracle and therefore we cannot open a Service Request. We will place an order for new batteries but it will take two weeks to arrive. In the meantime the performance is sluggish and we have a lot of complaints from the users. Is there any way to force the cache to be enabled manually, maybe from command line, until the batteries arrive? We have a very big problem here!

Thanks

Dear Vedha,

  unfortunately I found that since the storage is too old we don't have a maintenance contract active with Oracle and therefore we cannot open a Service Request. We will place an order for new batteries but it will take two weeks to arrive. In the meantime the performance is sluggish and we have a lot of complaints from the users. Is there any way to force the cache to be enabled manually, maybe from command line, until the batteries arrive? We have a very big problem here!

Thanks

Hello,

It seems that you did what was needed but if the cache remains disabled, this means that there is likely another issue and we cannot make assumptions without diagnostic data.

I may certainly have another proposal but I need to check this tomorrow morning from 09:00 CEST. I will provide another update tomorrow.

Regards

Nicolas

Dear Nicolas,

thank you very much! I hope we can find a temporary solution for this until the batteries arrive! Let me know if there is any way I can send you any data that will help you with the analysis.

Thanks,

Savvas

Hello,

What exact array model is it ?

Thanks

Nicolas

Nicolas,

  sorry for the late reply. I was at a meeting. It's SUN STORAGE 2510 ISCSI.

Thanks

Hello,

Thanks.

I would suggest to follow the action plan below:

1. Execute the following CAM commands to disable the battery expiration timer:

1. service -d <arrayname> -c set -q nvsram region=0xEE offset=0x2D value=0xFF

2. service -d <arrayname> -c set -q nvsram region=0xEE offset=0x2E value=0xFF

/* `service` is under:

/* Solaris: /opt/SUNWsefms/bin/

/* Linux: /opt/sun/cam/private/fms/bin/

/* Windows: C:\Program Files\Sun\Common Array Manager\Component\fms\bin\

2. Wait 10 min.

3. Check within CAM if the cache is re-enabled and check also the battery status.

Regards

Nicolas

Dear Nicolas,

where can i find the array name. In the screenshot below, Is it "ESXi_storage"?

Thanks.

Dear Nicolas,

this is the result I got from running the commands:

C:\Program Files\Sun\Common Array Manager\Component\fms\bin>service -d ESXi_storage -c set -q nvsram region=0xEE offset=0x2D value=0xFF

Executing the set command on ESXi_storage

Completion Status: ERROR: unable to open a connection to controller B

C:\Program Files\Sun\Common Array Manager\Component\fms\bin>service -d ESXi_storage -c set -q nvsram region=0xEE offset=0x2E value=0xFF

Executing the set command on ESXi_storage

Completion Status: ERROR: unable to open a connection to controller B

Please advice.

Thanks

Hello,

I will answer here as Nicolas seem to be just busy.

It seems that you have no working network connection to ctrl.B and therefore above command cannot be excuted. For this a network connection to both controllers is necesssary.

This could also be the problem for the cache issue as it remains disabled

First you have to check your network connection to ctrl.B As you mentioned already earlier that both controllers were rebooted I think there should be no issues with a hung network on the controllerside.

Perhaps a HW issue or config issue for ctrl.B network.

There is also a KM Doc available with some troubleshooting steps for network issues:

How to validate Sun Storage 6000, 2500 and Flexline Array Controller Out of Band Communication (Doc ID 1008327.1)

best regards

Uwe

Dear Uew,

thanks for your reply. We will test this and revert.

Thanks,

Savvas

Dear Uwe,

  here is what we tried:

Step A:

C:\Program Files\Sun\Common Array Manager\Component\fms\bin>service -d ESXi_storage -c contact

Executing the contact command on ESXi_storage

Attempting to contact the array using the following address(es):

192.168.4.205

192.168.4.206

Controller A is accessible via:

192.168.4.205 (oob)

Controller B is not accessible

Completion Status: Finished

Step B:

I tried to ping both IPs and everything is ok.

Step C:

I tried to telnet to 4.205 on 2463 and it was successful. 4.206 is not answering.

C:\Documents and Settings\Administrator>telnet 192.168.4.206 2463

Connecting To 192.168.4.206...Could not open connection to the host, on port 2463: Connect failed

How do you suggest we should continue troubleshooting this?

Thanks,

Savvas

Hello,

You have some communication issues with the controller B and you may have other technical issues with it. At this stage this goes beyond what we can really cover in the community. I am afraid that for further in depth assistance it will be required to have a SR.

Please contact your local support sales representative.

Regards

Nicolas

Dear Nicolas,

thank you very much for your effort. Just one last question. From your experience, will we have to change the controller or it might be a configuration issue that suddenly went bad?

Thanks,

Savvas

Hello,

I’m afraid it won’t be possible to answer that question without solid diagnostic data. Any answer would be just pure speculation and would not necessarily apply to your situation.

Regards

Nicolas

Dear Nicolas,

  thanks for your help. Really appreciated!

Thanks,

Savvas

Hello,

perhaps one last suggestion/hint if not already done to exclude a problem on your network.

You can install CAM on a Laptop and direct connect your Laptop to array controller to check if this works.

best regards

Uwe

Dear Uwe,

  thanks for the email. The problem is that the storage is at one of our remote offices and we have nobody there who can do this. What I did was to ask them to exchange the cables from the two management port to see if there was any switch or cable problem. The problem persists. Controller A pings and telnets without problems but Controller B does not.

Regards,

Savvas