Re: ISM corruption

On 8 June, 15:56, Jeremiah Leong <yogensha4e...@gmail.com> wrote: > How does one go about and fix the ISM if it gets corrupted after a > server crash? > > I've read my predecessor's notes and he's only written the easy way. > He said it's not exactly the correct way. The steps involve renaming > the index and mm folders and restarting ISM. Then he goes on to create > the backup devices and runs a backup. On his notes, he said the > previous backups become unusable but he also said that there are steps > to make them usable. Sadly, he didn't write them down. 1. http://publib.boulder.ibm.com/infocenter/idshelp/v111/index.jsp?topic=/com.ibm.ism.doc/ism42.htm to repair indexes. 2. Run nsr_admin -c Step thorough and if any devices are disabled re-enable them. 3. Run ism_show -devices ism_show -volumes to show what devices and volumes ISM knows about. 4, If any volumes are missing then either insert the tape for the volume into that device or link the volume directory to that device. ism_op -mount <device> to mount the device. ism_catalog -recreate from <device> 5. Worst case is restore from bootstrap volume: ism_catalog -find_bootstrap <device> ism_catalog -recover then rerun step 4 to recatalog any missing volumes. Hopefully you have been running "$INFORMIXDIR/bin/savegrp -O -l full <poolname> to save full bootstrap info after each backup! then