In the temporary directory you should locate and save the _raw.maegz file which contains the Glide results up to the point when your job crashed. Then, open the .log file and see what is the ligand number when your job stopped. Rename the .in file and open it. Change the LIGAND_START parameter to the ligand number that you experienced the crash.
Another way to recover the job is to list the active (and unfinished) jobs with "jobcontrol".
As mentioned in the jobcontrol manual you can type
"$SCHRODINGER/jobcontrol -list" to list all the active jobs and
"$SCHRODINGER/jobcontrol -recover <jobid>" to recover your job, which means that the _raw.maegz file will be simply copied in the original working directory.
The last option, which I have not tried is to restart the Glide job with the
"$SCHRODINGER/glide -restart" option, which however works only for distributed glide jobs (as mentioned in the manual).