Unix4life

SVM - Submirrors - One in Needs maintenance & another in Last Erred Main Point to Note: Always replace components in the “Maintenance” state first, followed by those in the “Last Erred” state. After a component is replaced and resynchronized, use the  metastat  command to verify its state. Then, validate the data. Theory from Oracle: When a component in a RAID-1 or RAID-5 volume experiences errors, Solaris Volume Manager puts the component in the “Maintenance” state. No further reads or writes are performed to a component in the “Maintenance” state. Sometimes a component goes into a “Last Erred” state. For a RAID-1 volume, this usually occurs with a one-sided mirror. The volume experiences errors. However, there are no redundant components to read from. For a RAID-5 volume this occurs after one component goes into “Maintenance” state, and another component fails. The second component to fail goes into the “Last Erred” state. When either a RAID-1 volume or...

UPGRADING  POWERPATH FROM EARLIER VERSIONS: Download the latest version of the EMC Grab utilities  available on EMC Online Support. Run PowerPath Configuration Checker. (Note: When you upgrade PowerPath, customized PowerPath settings are preserved (except write throttle queue settings) You must remove the earlier version of PowerPath before upgrading to the latest PowerPath version. When you remove PowerPath, follow the instructions in the documentation for that version. (Note: however, that you do not need to reboot the host after you remove the earlier version) (Note: In a cluster, you must follow the cluster upgrade procedure, which includes booting the host out of the cluster before you upgrade.) (The upgrade preserves your existing PowerPath license. You do not need to retype the PowerPath license information. ) You must shut down all applications before you upgrade PowerPath. If the host boots off a PowerPath pseudo device, you must reboot after t...

Notes: Use Below commands to replace a failed disk controlled by Solaris Volume Manager. Ensure to check the Logical Names of the disks ( Cxdxdx ) from format command. Below commands are for reference only. They should be followed at your own discretion. 1. df -k ( check the mounted file system associated with the disk) 2. format (check the status of the failed drive - showing unavailable) 3. metastat (check the meta disk sets/mirrors that the failed drive participated in) 4. metadb (check the metadb copies and their slice location of the failed drive) 5. metadb -d c0t0d0s7 (remove the metadb copies of the drive) 6. metadb (verify if the copies are gone) 7. metadetach d6 d16 (break mirror for disk set d6 mirror 16 on the failed drive) 8. metastat d6 (verify d6 status after breaking the mirror) 9. metadetach -f d6 d16 (break mirror by force) 10.metastat d5 11.metadetach -f d5 d15 (break mirror in d5) 12.metastat d1 13.metadetach -f d1 d11 (break mirror in d1) 14.met...