virgo0880
asked on
Orphaned Filespaces on TSM Server
Hi All,
I want to free up some of the space on my TSM server. So, I was wondering - is there a way to find out orphaned filespaces that can be deleted so that some of the tapes/storage can be freed by doing this. Is there any query using which I can find out the old filespaces and then I can decide which are required and which are not? I know i have to consider our company's retention policies, but first I want to come up with that list. Let me know how this can be done or any suggestions on doing this?
Thanks
Virgo
I want to free up some of the space on my TSM server. So, I was wondering - is there a way to find out orphaned filespaces that can be deleted so that some of the tapes/storage can be freed by doing this. Is there any query using which I can find out the old filespaces and then I can decide which are required and which are not? I know i have to consider our company's retention policies, but first I want to come up with that list. Let me know how this can be done or any suggestions on doing this?
Thanks
Virgo
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
If I run that query, whether it will put any load on the TSM server? Is it ok to run that query during the time backups are running or any migration or copy pool processes are running?
It's just a tiny query, no joins, no grouping, no aggregates, just a sort.
Depending on the amount of nodes and filespaces it will probably take less than two seconds.
And it doesn't interfere with any kind of server processing.
Depending on the amount of nodes and filespaces it will probably take less than two seconds.
And it doesn't interfere with any kind of server processing.
ASKER
I am getting following error while running that query :
ANR2964E The reference 'NODE_NAME' is not a column of the SQL result table.
|
..........................
ays(current_timestamp)-day
Also, removed order by statement and it says nothing to show. Can you explain that query what each query in that statement means?
Thanks
ANR2964E The reference 'NODE_NAME' is not a column of the SQL result table.
|
..........................
ays(current_timestamp)-day
Also, removed order by statement and it says nothing to show. Can you explain that query what each query in that statement means?
Thanks
SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
This is the query I am running from TSM GUI console:
select cast(NODE_NAME as char(18)) NODE, cast(FILESPACE_NAME as char(32)) as "FSNAME ", cast(FILESPACE_ID as char(3)) FSID, cast(BACKUP_START as char(16)) START, cast(BACKUP_END as char(16)) END, cast(days(current_timestam
Also, I want to find out how much data is stalled on my TSM server for these nodes? Is there any query for that?
Thanks
select cast(NODE_NAME as char(18)) NODE, cast(FILESPACE_NAME as char(32)) as "FSNAME ", cast(FILESPACE_ID as char(3)) FSID, cast(BACKUP_START as char(16)) START, cast(BACKUP_END as char(16)) END, cast(days(current_timestam
Also, I want to find out how much data is stalled on my TSM server for these nodes? Is there any query for that?
Thanks
Your query looks good so far. Why do you put everything into one line?
I don't know much about the GUI, but aren't there macros or scripts?
If you run the above thing from the GUI - I guess that you can't specify command line parameters?
Entering such a parameter is mandatory with this script however, see my first comment.
So I assume that the GUI treats "$1" verbatim as the node name to search for.
Since you probably don't have such a node you won't get any results, of course.
If you can't create a script omit this selector completely, so -
...
... cast(days(current_timestam
What do you mean with "stalled" data? I never heard that term before.
I don't know much about the GUI, but aren't there macros or scripts?
If you run the above thing from the GUI - I guess that you can't specify command line parameters?
Entering such a parameter is mandatory with this script however, see my first comment.
So I assume that the GUI treats "$1" verbatim as the node name to search for.
Since you probably don't have such a node you won't get any results, of course.
If you can't create a script omit this selector completely, so -
...
... cast(days(current_timestam
What do you mean with "stalled" data? I never heard that term before.
ASKER
Actually, I have taken out the list of nodes which are not active since more than a year and when I see "query filesp" for that node, I see filespaces which are lying there and the sizes for this filespaces are in GBs and TBs, so I was thinking if using query I can find out total size used by that node and delete that filespaces to reclaim that storage. Whether that make sense?
Also in your previous comments, you are talking about TSM scripts. Can you guide me on how to create that sql query script in TSM?
Thanks
Also in your previous comments, you are talking about TSM scripts. Can you guide me on how to create that sql query script in TSM?
Thanks
If you are willing to go to the command line, yes.
ASKER
Can you provide query for that ?
Also how do I create a tsm script for the query?
Also how do I create a tsm script for the query?
SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
Audit occupancy helps. I can easily generate the report using that command now.
Thanks
Thanks
ASKER
I created the script and tried, but getting following error:
tsm: TSM>query script query1 f=l
Name Line Command
Number
---------- ------ --------------------------
QUERY1 1 select cast(NODE_NAME as char(18)) NODE, -
6 cast(FILESPACE_NAME as char(32)) as "FSNAME ", -
11 cast(FILESPACE_ID as char(3)) FSID, -
16 cast(BACKUP_START as char(16)) START, -
21 cast(BACKUP_END as char(16)) END, -
26 cast(days(current_timestam
DAYS_AGO -
31 from filespaces where NODE_NAME like upper('%$1%') -
36 and days(current_timestamp)-da
41 order by NODE_NAME, 5
tsm: TSM>
tsm: TSM>run query1
ANR1464E RUN: Command script QUERY1, line 1, parameter 1 was not specified: select cast(NODE_NAME as char(18)) NODE, cast(FILESPACE_NAME as
char(32)) as "FSNAME ", cast(FILESPACE_ID as char(3)) FSID, cast(BACKUP_START as char(16)) START, cast(BACKUP_END as char(16)) END,
cast(days(current_timestam
days(current_timestamp)-da
ANR1463E RUN: Command script QUERY1 completed in error.
ANS8001I Return code 3.
also tried specifying node name as an argument, but different error :
tsm: TSM>run query1 node1
ANR2964E The reference 'NODE_NAME' is not a column of the SQL result table.
|
..........................
ays(current_timestamp)-day
ANR1463E RUN: Command script QUERY1 completed in error.
ANS8001I Return code 3.
tsm: TSM>query script query1 f=l
Name Line Command
Number
---------- ------ --------------------------
QUERY1 1 select cast(NODE_NAME as char(18)) NODE, -
6 cast(FILESPACE_NAME as char(32)) as "FSNAME ", -
11 cast(FILESPACE_ID as char(3)) FSID, -
16 cast(BACKUP_START as char(16)) START, -
21 cast(BACKUP_END as char(16)) END, -
26 cast(days(current_timestam
DAYS_AGO -
31 from filespaces where NODE_NAME like upper('%$1%') -
36 and days(current_timestamp)-da
41 order by NODE_NAME, 5
tsm: TSM>
tsm: TSM>run query1
ANR1464E RUN: Command script QUERY1, line 1, parameter 1 was not specified: select cast(NODE_NAME as char(18)) NODE, cast(FILESPACE_NAME as
char(32)) as "FSNAME ", cast(FILESPACE_ID as char(3)) FSID, cast(BACKUP_START as char(16)) START, cast(BACKUP_END as char(16)) END,
cast(days(current_timestam
days(current_timestamp)-da
ANR1463E RUN: Command script QUERY1 completed in error.
ANS8001I Return code 3.
also tried specifying node name as an argument, but different error :
tsm: TSM>run query1 node1
ANR2964E The reference 'NODE_NAME' is not a column of the SQL result table.
|
..........................
ays(current_timestamp)-day
ANR1463E RUN: Command script QUERY1 completed in error.
ANS8001I Return code 3.
As for the parameter - I told you several times that specifying a parameter
is mandatory with this script!
Specify "%" for all nodes.
I have no explanation for the second error, the script works for me.
Anyway, replacing NODE_NAME with "1" in the order clause should help.
Run this on the dsmadmc command line:
upd scr query1 "order by 1, 5" line=41
and retry (with a parameter!)
is mandatory with this script!
Specify "%" for all nodes.
I have no explanation for the second error, the script works for me.
Anyway, replacing NODE_NAME with "1" in the order clause should help.
Run this on the dsmadmc command line:
upd scr query1 "order by 1, 5" line=41
and retry (with a parameter!)
ASKER
Thanks that worked.
One more thing - what is the difference between "query auditoccupancy" and "query filespaces" as I see there is a difference in space utilization for the nodes using both commands. Query auditocc shows some nodes higher in space utilization while some lower. Also, I used "query filespace nodename" and added the space utilization for all the filespaces from each node. Why there is a difference in number between these two commands? do you have any idea about it?
Thanks
One more thing - what is the difference between "query auditoccupancy" and "query filespaces" as I see there is a difference in space utilization for the nodes using both commands. Query auditocc shows some nodes higher in space utilization while some lower. Also, I used "query filespace nodename" and added the space utilization for all the filespaces from each node. Why there is a difference in number between these two commands? do you have any idea about it?
Thanks
Q Filespaces shows the volumes making up client storage (capacity and used percentage) which are eligible for backup as seen by the TSM client and reported to the TSM server. Filespaces as such don't occupy any TSM server space.
Q Occupancy shows the amount of TSM server storage actually used for backing up these filespaces.
Occupancy of a given client is usually much higher than the sum of the client's used filespace capacity due to versioning.
Q Occupancy shows the amount of TSM server storage actually used for backing up these filespaces.
Occupancy of a given client is usually much higher than the sum of the client's used filespace capacity due to versioning.
ASKER
So, if I want to free up that space for the nodes which are not active since many days. I can just use "del filespace nodename *" and get that space back or is there a different command for deleting the occupancy of the filespaces on TSM server storage?
I thought that "query filespaces" will show the space used by node on TSM Server and using "del filespace" I can reclaim that space.. Whether this is true?
Thanks
I thought that "query filespaces" will show the space used by node on TSM Server and using "del filespace" I can reclaim that space.. Whether this is true?
Thanks
DEL FILESPACE is quite correct to delete backup/archive data from server storage.
Although a "filespace" is basically client storage (not server storage) TSM organizes its server storage under a structure derived from and related to the client's filespace names.
"DEL FILESPACE" is a bit misleading, I admit, because this command does not delete client storage but backed up client data in TSM server storage.
Regard "DEL FILESPACE" as a short form for "delete backed up filespace data" (this latter is not a command, it's my own "pseudo code"!)
Although a "filespace" is basically client storage (not server storage) TSM organizes its server storage under a structure derived from and related to the client's filespace names.
"DEL FILESPACE" is a bit misleading, I admit, because this command does not delete client storage but backed up client data in TSM server storage.
Regard "DEL FILESPACE" as a short form for "delete backed up filespace data" (this latter is not a command, it's my own "pseudo code"!)
ASKER
Got it, one last question. I have one node which shows like 11 to 12 TB of data on TSM tape storage pool and I have got a go ahead for deleting and removing the filespaces for that node. When I do "query nodedata nodename" for that node, I see a lot of tapes being used by that node. So if I remove the data for that node and the node itself (by deleting all the filespaces of that node), whether I will get all that tapes back as "Scratch", is that right?
I have attached the output of "query nodedata" command for that node.
Thanks
I have attached the output of "query nodedata" command for that node.
Volume Name Storage Device Estimated Pct Volume
Pool Name Class Name Capacity Util Status
------------------------ ----------- ---------- --------- ----- --------
005013 TBACK2 TDEV3592 626,941.0 35.6 Full
005026 TBACK2 TDEV3592 272,308.8 76.3 Full
005028 TBACK2 TDEV3592 325,466.5 99.1 Full
005031 TBACK2 TDEV3592 645,871.3 47.8 Full
005045 TBACK2 TDEV3592 336,583.9 76.3 Full
005068 TBACK2 TDEV3592 446,623.2 36.5 Full
005080 TBACK2 TDEV3592 270,235.4 92.8 Full
005086 TBACK2 TDEV3592 413,262.6 59.1 Full
005094 TBACK2 TDEV3592 321,135.3 50.8 Full
005097 TBACK2 TDEV3592 348,515.2 83.5 Full
005135 TBACK2 TDEV3592 419,104.2 54.3 Full
005141 TBACK2 TDEV3592 312,462.2 37.1 Full
005145 TBACK2 TDEV3592 521,453.2 40.4 Full
005181 TBACK2 TDEV3592 900,000.0 26.4 Filling
005184 TBACK2 TDEV3592 294,067.9 69.2 Full
005186 TBACK2 TDEV3592 322,946.8 90.0 Full
005217 TBACK2 TDEV3592 446,572.7 73.6 Full
005235 TBACK2 TDEV3592 310,758.2 91.4 Full
005310 TBACK2 TDEV3592 328,658.2 99.8 Full
005316 TBACK2 TDEV3592 340,735.5 83.2 Full
005324 TBACK2 TDEV3592 331,246.7 61.4 Full
005337 TBACK2 TDEV3592 319,216.8 93.0 Full
005353 TBACK2 TDEV3592 876,436.0 36.2 Full
005358 TBACK2 TDEV3592 378,829.9 78.4 Full
005370 TBACK2 TDEV3592 320,585.1 38.1 Full
005373 TBACK2 TDEV3592 409,968.7 84.6 Full
005377 TBACK2 TDEV3592 561,882.7 54.2 Full
005384 TBACK2 TDEV3592 598,458.2 38.1 Full
005413 TBACK2 TDEV3592 467,805.3 47.8 Full
005426 TBACK2 TDEV3592 327,788.9 40.3 Full
005436 TBACK2 TDEV3592 429,058.1 77.6 Full
005456 TBACK2 TDEV3592 389,305.3 37.5 Full
005461 TBACK2 TDEV3592 310,631.8 84.5 Full
005467 TBACK2 TDEV3592 505,025.0 32.9 Full
005468 TBACK2 TDEV3592 397,278.1 96.6 Full
005472 TBACK2 TDEV3592 336,249.6 74.6 Full
005477 TBACK2 TDEV3592 300,855.5 61.4 Full
005496 TBACK2 TDEV3592 390,289.7 33.9 Full
005501 TBACK2 TDEV3592 329,405.4 98.0 Full
005521 TBACK2 TDEV3592 339,700.4 98.3 Full
005566 TBACK2 TDEV3592 363,287.5 100.0 Full
005574 TBACK2 TDEV3592 320,891.3 70.9 Full
005597 TBACK2 TDEV3592 354,104.2 95.0 Full
005620 TBACK2 TDEV3592 310,130.0 84.0 Full
005646 TBACK2 TDEV3592 336,251.8 85.8 Full
005686 TBACK2 TDEV3592 461,031.5 81.3 Full
005689 TBACK2 TDEV3592 350,911.4 64.9 Full
005739 TBACK2 TDEV3592 532,548.3 44.4 Full
005756 TBACK2 TDEV3592 292,996.7 71.8 Full
005773 TBACK2 TDEV3592 470,994.7 98.7 Full
005779 TBACK2 TDEV3592 326,306.4 96.9 Full
005806 TBACK2 TDEV3592 419,841.0 60.5 Full
005819 TBACK2 TDEV3592 358,584.3 41.1 Full
005821 TBACK2 TDEV3592 318,441.9 87.8 Full
005840 TBACK2 TDEV3592 331,333.9 66.2 Full
005858 TBACK2 TDEV3592 364,738.5 49.5 Full
005866 TBACK2 TDEV3592 573,260.0 51.8 Full
005880 TBACK2 TDEV3592 362,064.7 63.5 Full
005893 TBACK2 TDEV3592 307,101.3 68.3 Full
005904 TBACK2 TDEV3592 268,650.0 63.4 Full
005920 TBACK2 TDEV3592 335,009.0 77.8 Full
005927 TBACK2 TDEV3592 325,488.4 67.4 Full
005929 TBACK2 TDEV3592 243,218.3 31.7 Full
005935 TBACK2 TDEV3592 401,239.2 44.6 Full
005953 TBACK2 TDEV3592 320,726.3 72.7 Full
005958 TBACK2 TDEV3592 303,882.5 96.7 Full
005961 TBACK2 TDEV3592 296,214.7 48.8 Full
006005 TBACK2 TDEV3592 346,129.0 50.4 Full
006008 TBACK2 TDEV3592 432,490.8 32.7 Full
006031 TBACK2 TDEV3592 328,684.8 61.5 Full
006045 TBACK2 TDEV3592 478,331.8 68.6 Full
006122 TBACK2 TDEV3592 532,128.0 41.0 Full
006260 TBACK2 TDEV3592 363,472.5 100.0 Full
006261 TBACK2 TDEV3592 345,595.1 71.2 Full
006386 TBACK2 TDEV3592 900,000.0 19.1 Filling Thanks
Is this really "Q NODEDATA" output?
Or is it rather "Q VOL"?
With Q NODEDATA you should see a column "Physical space occupied" and that's the space you will get back after "DEL FILESPACE".
If a tape volume is full with only that node's data it will become scratch, if there is also data of other nodes on it then it will obviously not become scratch, it will just become "emptier", or "less full" if you prefer.
Btw. the space will be freed immediately, you won't have to wait for the next "Expire Inventory" run.
Or is it rather "Q VOL"?
With Q NODEDATA you should see a column "Physical space occupied" and that's the space you will get back after "DEL FILESPACE".
If a tape volume is full with only that node's data it will become scratch, if there is also data of other nodes on it then it will obviously not become scratch, it will just become "emptier", or "less full" if you prefer.
Btw. the space will be freed immediately, you won't have to wait for the next "Expire Inventory" run.
ASKER
Ohh sorry that was the output of q vol. I have attached the output of q nodedata for that node. The output is showing tapes from Primary as well as Copy storage pools, so for primary the space will be available immediately,
but what about Copy Storage pool tapes, that will be freed when the reclamation will start or expire inventory will start ?
but what about Copy Storage pool tapes, that will be freed when the reclamation will start or expire inventory will start ?
file2node 005013 TBACK2 19,629.4
3
file2node 005026 TBACK2 207,646.
22
file2node 005028 TBACK2 322,465.
00
file2node 005031 TBACK2 38,515.8
4
file2node 005037 CBACK2 149,500.
43
file2node 005057 CBACK2 189,978.
81
file2node 005058 CBACK2 187,484.
31
file2node 005068 TBACK2 641.72
file2node 005086 TBACK2 20,868.3
6
file2node 005094 TBACK2 163,261.
56
file2node 005125 CBACK2 330,304.
33
file2node 005135 TBACK2 4,204.52
file2node 005149 CBACK2 277,937.
61
file2node 005166 CBACK2 3,477.83
file2node 005217 TBACK2 325,961.
47
file2node 005220 CBACK2 223,455.
15
file2node 005263 CBACK2 199,027.
20
file2node 005276 CBACK2 153.76
file2node 005281 CBACK2 183,345.
09
file2node 005301 CBACK2 264,185.
51
file2node 005310 TBACK2 327,991.
05
file2node 005316 TBACK2 283,650.
83
file2node 005329 CBACK2 280,464.
29
file2node 005337 TBACK2 296,940.
84
file2node 005353 TBACK2 316,931.
41
file2node 005358 TBACK2 259,937.
20
file2node 005370 TBACK2 122,067.
40
file2node 005373 TBACK2 346,707.
84
file2node 005377 TBACK2 26,482.6
1
file2node 005380 CBACK2 208,573.
73
file2node 005384 TBACK2 228,011.
16
file2node 005413 TBACK2 38,108.6
0
file2node 005418 CBACK2 79,497.4
8
file2node 005426 TBACK2 132,183.
66
file2node 005431 CBACK2 269,875.
93
file2node 005444 CBACK2 206,897.
69
file2node 005456 TBACK2 145,928.
68
file2node 005467 TBACK2 166,339.
60
file2node 005468 TBACK2 213,197.
92
file2node 005477 TBACK2 184,666.
63
file2node 005501 TBACK2 322,832.
22
file2node 005521 TBACK2 333,952.
64
file2node 005528 CBACK2 119,713.
07
file2node 005550 CBACK2 182,398.
87
file2node 005555 CBACK2 198,265.
72
file2node 005566 TBACK2 231,840.
29
file2node 005574 TBACK2 227,616.
20
file2node 005584 CBACK2 20,347.3
9
file2node 005597 TBACK2 204,894.
01
file2node 005601 CBACK2 344,542.
65
file2node 005608 CBACK2 271,234.
94
file2node 005646 TBACK2 288,654.
91
file2node 005657 CBACK2 38,472.1
1
file2node 005678 CBACK2 38,108.6
0
file2node 005683 CBACK2 26,941.7
8
file2node 005686 TBACK2 374,704.
39
file2node 005713 CBACK2 186,645.
60
file2node 005739 TBACK2 92,271.6
7
file2node 005752 CBACK2 244,646.
77
file2node 005756 TBACK2 210,330.
60
file2node 005763 CBACK2 350.09
file2node 005764 CBACK2 246,892.
48
file2node 005773 TBACK2 335,948.
13
file2node 005776 CBACK2 308,768.
65
file2node 005779 TBACK2 316,104.
01
file2node 005781 CBACK2 444,562.
90
file2node 005790 CBACK2 295,749.
01
file2node 005794 CBACK2 299,298.
52
file2node 005799 CBACK2 337,468.
47
file2node 005821 TBACK2 279,695.
22
file2node 005832 CBACK2 214,956.
51
file2node 005840 TBACK2 219,384.
37
file2node 005843 CBACK2 332,388.
01
file2node 005845 CBACK2 226,461.
32
file2node 005857 CBACK2 249,282.
72
file2node 005866 TBACK2 297,131.
84
file2node 005880 TBACK2 229,745.
67
file2node 005886 CBACK2 102,240.
49
file2node 005889 CBACK2 336,383.
30
file2node 005893 TBACK2 209,813.
38
file2node 005920 TBACK2 260,784.
55
file2node 005927 TBACK2 219,297.
27
file2node 005929 TBACK2 77,188.1
8
file2node 005933 CBACK2 283,045.
47
file2node 005935 TBACK2 179,021.
46
file2node 005944 CBACK2 189,914.
92
file2node 005947 CBACK2 148,088.
10
file2node 005953 TBACK2 233,141.
96
file2node 005958 TBACK2 293,996.
59
file2node 005960 CBACK2 25,498.5
3
file2node 005979 CBACK2 176,054.
04
file2node 005980 CBACK2 20,868.3
6
file2node 006005 TBACK2 174,491.
87
file2node 006008 TBACK2 78,549.6
0
file2node 006031 TBACK2 202,207.
10
file2node 006045 TBACK2 325,986.
32
file2node 006061 CBACK2 176,005.
29
file2node 006122 TBACK2 218,433.
38
file2node 006139 CBACK2 675.39
file2node 006190 CBACK2 106.52
file2node 006193 CBACK2 273,818.
54
file2node 006237 CBACK2 118,442.
66
file2node 006245 CBACK2 351,378.
32
file2node 006260 TBACK2 325,943.
36
file2node 006261 TBACK2 206,947.
84
file2node 006275 CBACK2 348,824.
84
file2node 006276 CBACK2 295,690.
71
file2node 006298 CBACK2 444,015.
88
file2node 006315 CBACK2 228,493.
94
file2node 006386 TBACK2 55,185.2
3
It will also be freed immediately.
Of course for tapes only partially freed (primary or copy doesn't matter) a space reclamation will take place (if the threshold is reached), just like after an "expire inventory" run.
Of course for tapes only partially freed (primary or copy doesn't matter) a space reclamation will take place (if the threshold is reached), just like after an "expire inventory" run.
ASKER
Thanks for the explanation.
ASKER
Is it possible to see what all tapes got freed up by deleting the filespaces of the node or any way of tracking that?
dsmadmc -id=userid -pass=password -console
Watch the console output and look for
Scratch volume volser has been deleted from storage pool poolname
You can issue Q PR regularly and watch the "DELETE FILESPACE" process, telling you how many items already got deleted.
Or run
select count(*) from libvolumes where STATUS='Scratch'
regularly to see the number of scratch volumes grow.
Watch the console output and look for
Scratch volume volser has been deleted from storage pool poolname
You can issue Q PR regularly and watch the "DELETE FILESPACE" process, telling you how many items already got deleted.
Or run
select count(*) from libvolumes where STATUS='Scratch'
regularly to see the number of scratch volumes grow.
ASKER
Thanks much for your help on this.Giving points.
ASKER
But one thing I noticed that after clearing up all the filespaces and nodes, my db backup is taking half an hour more to what was it was taking before like before the db backup (snapshot and full backup) use to take 1.5 hours to complete but now it is taking 2.5 hours, even though there is a reduction in DB size as I deleted the filespaces and nodes. What can be the reason for that?
I can open a new thread for that. But just want to discuss this with you before that? Let me know if you want new thread for that?
Thanks
I can open a new thread for that. But just want to discuss this with you before that? Let me know if you want new thread for that?
Thanks
I fear I can't help you with this last issue.
I'm running TSM 6 for quite a while now, and this version sits on DB2 instead of the old proprietary thing.
DB2 is self-reorganizing, so there is never any trouble in that aspect.
I can imagine that the database pointers have to go long distances to get the backup in line, because of the gaps resulting from your big delete action.
I could further imagine that a database reorganization could help, but this would require a (perhaps long) downtime:
Here are the steps:
- Halt the server
- Issue DSMSERV UNLOADDB
- Issue DSMSERV LOADFORMAT
- Issue DSMSERV LOADDB
- Issue DSMSERV AUDITDB
Here are more details and the exact command formats:
http://publib.boulder.ibm.com/infocenter/tivihelp/v1r1/topic/com.ibm.itsmaixn.doc/anragd55606.htm?path=2_1_6_2_6_6#compdb1
As I wrote, it's only a guess and I don't know if reorgainizing the DB will indeed reduce backup times, but I assume so.
I'm running TSM 6 for quite a while now, and this version sits on DB2 instead of the old proprietary thing.
DB2 is self-reorganizing, so there is never any trouble in that aspect.
I can imagine that the database pointers have to go long distances to get the backup in line, because of the gaps resulting from your big delete action.
I could further imagine that a database reorganization could help, but this would require a (perhaps long) downtime:
Here are the steps:
- Halt the server
- Issue DSMSERV UNLOADDB
- Issue DSMSERV LOADFORMAT
- Issue DSMSERV LOADDB
- Issue DSMSERV AUDITDB
Here are more details and the exact command formats:
http://publib.boulder.ibm.com/infocenter/tivihelp/v1r1/topic/com.ibm.itsmaixn.doc/anragd55606.htm?path=2_1_6_2_6_6#compdb1
As I wrote, it's only a guess and I don't know if reorgainizing the DB will indeed reduce backup times, but I assume so.
ASKER
Thanks for pointing out in that direction. I was thinking the same way. I will look into that.
ASKER