EMC storage systems need their software updated from time to time, and the way you do that is to copy the new software to the system and then apply the update.
However, the amount of hard drive space allocated to the system is limited, so it’s very common to run out of space. The fix is what you’d expect (you need to delete old files) and cleanup the system storage. Here’s how to do that:
- Enable remote access via SSH (PUTTY) and connect to the storage
- Login to Unisphere, goto Settings, Service System and enter your password. Under Service Actions run the Enable SSH option
- Connect to the system using PUTTY and enter in the proper username and password (Username is ‘service’ and password is the password you entered in step 1.)
- At this point if you’re being helped by Support they will run “svc_diag” to see if there are any active faults. If there’s none then they will enable the svc_service_shell service by running svc_inject -k and entering in the appropriate response code for the challenge code displayed.
- They then run “set_banner” and “PATH=$PATH:/nas/bin” and “crm_min -1nf”
- “df -kh” will show you the filesystem folders. The “11” command shows you the files with their names and sizes.
- Directory /cores/service contains the uploaded Software updates ending in .bin and .bin.tmpdir. You can remove them with the command “rm -rf <FILENAME>”
- Directory /EMC/backend/service/data_collection has old .tar files. You can remove the older ones and keep the newest.
- Directory /EMC/backend/service/data_collection/cores has old _dir files which you can remove as well, but they don’t take up that much space…
- Directory /var/tmp has a folder in it called upgrade which can be removed. That will clear any candidate update which may have been left loaded in Unisphere.
- Run “cf_ha_api.pl restart ECOM” to refresh the system.
- To see if your system is ready for an upgrade, run “check_health -pre” and look for PASS on each item.
- re-upload the Software update and then apply it.