asked on

Appropriate plan for maintaining DC

Dear Experts, we are planning to test the electricity in the next few days, so at that time there is no any power, even from the Generators.

So we are going to turn off our system at night and turn it on in the next morning, but how are the proper procedures?

We have:
1.Physical Cisco routers and switches, NAS storage
2.Physical IBM servers
3.Physical Sonicwall firewall
4.Physical ESXi hosts HPE Gen9
5.Physical HPE 8/8 Base 8-port Enabled SAN Switch AM867C, SAN storage HPE MSA2040

For #1,2,3 we can save the running configuration and shutdown without any problem, but #4 and #5

We got this plan from the Internet and it is quite understandable. Should we consider something else?

The order you will need to shute down is: 
    Shut down the VMs
    Place the hosts in MAintenance mode
    Shut down the hosts
    Shut down vCenter (if vCenter is a vm you will need to connect to each host seperately and shut it down.)

To bring everything back up
    Start the hosts
    Start vCenter

Open in new window

However, for SAN switch and SAN storage, we could not find the save config button anywhere. So could you please suggest the plan?

Many thanks as always!

ASKER CERTIFIED SOLUTION

John

membership

This solution is only available to members.

To access this solution, you must be a member of Experts Exchange.

Start Free Trial

kevinhsieh

Your networking equipment can have unsaved changes. For Cisco, do a "write memory" from a privileged session. You can make a backup of the configuration by saving the output from "show run".

For other vendors, look at their documentation.

Your SAN should have a shutdown command to safely and cleanly shut it down. Check with your vendor or documentation.

kevinhsieh

Sorry, I just reread your question. I have no experience with that particular family of switches or storage.

I would give HPE Support a call for the best support and to get the "warm fuzzies" that come with knowing that you understand all of the correct steps from the experts.

Sajid Shaik M

do you have any UPS ..in your environment ?

if yes in UPS it's an feature...with wich you can shutdowns and restart servers ... by using it's scripts.

all the best

Member_2_231077

No need to save the confug for the Brocade SAN switches as they are held in CMOS, however you can telnet to them and run configupload which wull FTP a config file to wherever you want.

The MSA has flash backed cache so will tolerate any length of power out, however you should shutdown both which you can do from the GUI. Saving the logs will save the config, https://support.hpe.com/hpsc/doc/public/display?docId=mmr_kc-0134161 , it is wrapped up in XML though so difficult to read. Obviously you wait for all VMs and anything else using the SAN to be shut down before shutting that down.

Michelangelo

Also, have a look at page 76-77 of
https://support.hpe.com/hpsc/doc/public/display?docId=c03792322

When a controller is shutdown or otherwise rendered inactive—its Link Status LED remains illuminated—falsely indicating that the controller can communicate with the host. Though a link exists between the host and the chip on the controller, the controller is not communicating with the chip. To reset the LED, the controller must be properly power-cycled [see “Powering on/powering off” (page 26)].
Cache Status LED details
If the LED is blinking evenly, a cache flush is in progress. When a controller module loses power and write cache contains data that has not been written to disk, the supercapacitor pack provides backup power to flush (copy) data from write cache to CompactFlash memory. When cache flush is complete, the cache transitions into self-refresh mode.
If the LED is blinking momentarily slowly, the cache is in a self-refresh mode. In self-refresh mode, if primary power is restored before the backup power is depleted (3–30 minutes, depending on various factors), the system boots, finds data preserved in cache, and writes it to disk. This means the system can be operational within 30 seconds, and before the typical host I/O time-out of 60 seconds, at which point system failure would cause host-application failure. If primary power is restored after the backup power is depleted, the system boots and restores data to cache from CompactFlash, which can take about 90 seconds. The cache flush and self-refresh mechanism is an important
data protection feature; essentially four copies of user data are preserved: one in controller cache and one in CompactFlash of each controller. The Cache Status LED illuminates solid green during the boot-up process. This behavior indicates the cache is logging all POSTs, which will be flushed to the CompactFlash the next time the controller shuts down.
CAUTION:
If the Cache Status LED illuminates solid green—and you wish to shut-down the controller—do so from the user interface, so unwritten data can be flushed to CompactFlash.