| Title: | Storage Library System |
| Moderator: | COOKIE::REUTER |
| Created: | Sun Oct 13 1991 |
| Last Modified: | Fri Jun 06 1997 |
| Last Successful Update: | Fri Jun 06 1997 |
| Number of topics: | 2270 |
| Total number of notes: | 7850 |
Okay,
I have a 4100 with 2 x TZ887's proposed. The customer want's to utilise up to
the 40GB / tape as it costs them
for the tapes. What I am trying to do is have multiple SBK's writing there
savesets to the same tape. Now
from scavenging around the notes and after loosing my long crafted source in
SLS (Tony Parsons) to OMS I need
the following answered :-
If I define, in the correct table, SYSBAK_DISM_QUALS to "/NOUNLOAD" in
multiple SBK's that are going to run
sequentially (maybe different nodes in a cluster and different times), will
this provide the facility to not
unload the tape say for those concurrent jobs so that the
savesets from these multiple SBK's appear on one tape or do I have to do
something else ?
The questions that confuse me are :-
1. The SBK's will have the CONTINUE string for those I wish to continue on the
same tape ... with I hear the
flag "/NEW" appended to the string in the first SBK.
2. The SYSBAK_DISM_QUALS is defined for the run of that SBK only, or is it
global, this screws me up with 2 TZ887's if this is true.
3. Do I have to deassign at the end of each SBK for these quals.
4. What happens to the label of the tape ( e.g if the tape is free and
allocated from the first backup , what happens
when the second backup runs to this as I assume the tape is now alloocated
from the first job , or is SLS just so cool
that it can handle this based on the CONTINUE statement.
Yours frustratedly
Craig Bowers
"Striving to keep the West going ... yee hah "
P.S I believe there used to be a Hitch Hikers or cookbook on SLS .. any
pointers to its location
so I can become better versed on this product.
[Posted by WWW Notes gateway]
| T.R | Title | User | Personal Name | Date | Lines |
|---|---|---|---|---|---|
| 2207.1 | CX3PST::BSS::SAUL | Tue Mar 04 1997 12:25 | 47 | ||
Hi Craig, >If I define, in the correct table, SYSBAK_DISM_QUALS to "/NOUNLOAD" in >multiple SBK's that are going to run >sequentially (maybe different nodes in a cluster and different times), will >this provide the facility to not >unload the tape say for those concurrent jobs so that the >savesets from these multiple SBK's appear on one tape or do I have to do >something else ? Assigning the logical will leave the tape in the drive, however if you are running in a robotic mode you probably don't want to do this. In a robotic mode, there is no need to use the logical since SLS will load whatever it needs from wherever it is in the magazine. So the first back finishes and moves the tape back to slot 0, the next backup starts using the CONTINUE and says get me that tape that is in slot 0 and load it. SLS will do this. >1. The SBK's will have the CONTINUE string for those I wish to continue on the >same tape ... with I hear the > flag "/NEW" appended to the string in the first SBK. This is true. You want to be sure to use NEXT_JOB to chain the SBKs together. If two jobs run with the same CONTINUE symbol set, the first to get the volume wins and the loser goes and gets a new volume. >2. The SYSBAK_DISM_QUALS is defined for the run of that SBK only, or is it >global, this screws me up with 2 TZ887's if this is true. You set up the logical on the LNM$SLS$VALUES table, so no matter when you assign it, its globally available to whoever is running at that time. >3. Do I have to deassign at the end of each SBK for these quals. Only if you no longer want them available globally. >4. What happens to the label of the tape ( e.g if the tape is free and >allocated from the first backup , what happens > when the second backup runs to this as I assume the tape is now alloocated >from the first job , or is SLS just so cool > that it can handle this based on the CONTINUE statement. Its cool...SLS opens up a continuation file to find the volume and asks for it to be loaded. Hope this helps... Ted | |||||
| 2207.2 | TZ8x7, I'd rather keep the tape loaded | HLSM01::HENK_H | Henk Hofman @UTO | Wed Mar 05 1997 18:51 | 20 |
RE .1 > Assigning the logical will leave the tape in the drive, however if you are > running in a robotic mode you probably don't want to do this. In a robotic > mode, there is no need to use the logical since SLS will load whatever it > needs from wherever it is in the magazine. So the first back finishes and > moves the tape back to slot 0, the next backup starts using the CONTINUE and > says get me that tape that is in slot 0 and load it. SLS will do this. Because Craig intends using TZ887's I'd rather leave the tape loaded. Unloading it will work but these robots do not win load/unload competitions. Initialising tapes in a TZ877 manually a 2 minutes wait-time is needed in between to consecutive initialize commands to allow for unloading a tape and loading the next one. I've not used a TZ887 yet but I assume the robotics are the same and it's just the drive that is different. Henk. | |||||
| 2207.3 | Only more Questions | NETRIX::"craig.bowers@mail.dec.com" | Craig Bowers | Wed Mar 05 1997 22:35 | 23 |
I thank you both for the responses ... its beginning to take shape now ...
but...
If you enter a CONT string in a SBK it creates a file in SLS$SYSBAK of
the format fname.CONT_READY associating the tape volume label with the
SBK's that u r writing to the same tape.
This is cool only if you now try and run the same backups to another tape
it prompts you that it is expecting the volume used and named previously
in the above file.
Due to the cycle I am setting up this is going to cause problems I suspect?
Can anyone enlighten me in to the relationship of the files created as a
result of using the CONT option so I can sus this out.
P.S There also seems to be a requirement for a fname.CONT_INPROG file ?
What is this ?
In anticipation
Craig
[Posted by WWW Notes gateway]
| |||||
| 2207.4 | PRSSOS::FONDI | Thu Mar 06 1997 10:57 | 11 | ||
>>>P.S There also seems to be a requirement for a fname.CONT_INPROG file ? >>> What is this ? all you want to know is... have a look at the file produce by $search/window=(1,70) sls$system:sysbak.com "Check out continuation reel" Hope this help Henri | |||||