I posed a question a couple of months about HSM dump time elongating. Our DASD geometry changed, so the number of devices changed. I followed what HSM was actually doing so I thought that I would pass it along in case anyone else would be interested. i hope that this isn't to convoluted, it almost makes sense to me.
HSM dumps the non SMS volumes after all of the SMS volumes have been dumped. Also HSM groups the DASD volumes into cluster of the DUMPCLASS' STACK increments. So in a scenario, we used two drives to dump 47 volumes, 11 of which were non SMS. HSM grouped the SMS volumes in three groups of 14, 14 and 8 (STACK parm was 14). As best as I can ascertain these group consists of StorGrp's but not in any order that I can determine. One of the SMS group of 14 ran longer than the other two groups but HSM wouldn't start the non SMS group until all three of the SMS groups completed and left me with 40 minutes of using only 1 tape drive. This caused my elongation. This also explained another problem. When a dump has an I/O error at the beginning and it can't be corrected, that group of volumes is retried later. In our scenario the two SMS groups of 14 start together and the SMS group of 8 start when either of the 2 groups of 14 finish. When the error occurs, as luck would have it, it's usually the longer (time) group of 14, so it waits unit the group of 8 is finished and then it starts. Jack Kelly 202-502-2390 (Office) ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
