OnDemand User Group
Support Forums => MP Server => Topic started by: danielito on December 14, 2022, 09:47:54 AM
-
Hi
I have another question about a slow System.
I have to GET documents out of an Applikation and load it again into a new one.
x.ind
x.out
x.res
On the Test System it takes less then 10 sec for 3 loads if I do it the same on the Prod it takes almost 9 min :o
Also if I take a look at this Documents it takes a long time till they show up ...
It's on a RHAT 8.0/Oracle and the performance was ok so far.
I am searching all over to find the reason .... no luck
If I have the ARS_NUM_DBSRVR=100 < Is this to high ?
d/V10.5/bin $ echo `date` arssockd count: `arssockd -I DB-p | wc -l`
arssockd count: 111
nd/V10.5/bin $ echo `date` arssockd count: `arssockd -I DB -p | wc -l`
arssockd count: 111
nd/V10.5/bin $ echo `date` arssockd count: `arssockd -I DB -p | wc -l`
arssockd count: 111
nd/V10.5/bin $ echo `date` arssockd count: `arssockd -I DB -p | wc -l`
arssockd count: 116
nd/V10.5/bin $ echo `date` arssockd count: `arssockd -I DB -p | wc -l`
arssockd count: 115
nd/V10.5/bin $ echo `date` arssockd count: `arssockd -I DB -p | wc -l`
arssockd count: 111
nd/V10.5/bin $ echo `date` arssockd count: `arssockd -I DB -p | wc -l`
arssockd count: 111
nd/V10.5/bin $ echo `date` arssockd count: `arssockd -I DB -p | wc -l`
Spectrum Protect is working and Storage S3 should also not be the problem because there where no changes done.
regards dani
-
100 is a really high number. How many users are typically accessing your system at a given time. Keep in mind the default is 4.
-
During the day there are normally 60 -115 but it can be more and with all the user there are still File loads running.
-
ARS_NUM_DBSRVR = 8 is probably all you need. The likely hood that you will have 100 users hit search at the exact same time is slim to none.
-
ARS_NUM_DBSRVR = 8 is probably all you need. The likely hood that you will have 100 users hit search at the exact same time is slim to none.
thanks aj I will try this .....
-
Do you have any custom code in place that could be causing these processes to be long lived?
Maybe a Security Exit or Query exit or permissions exit? If not done properly these could cause issues.
-
Ahh this is maybe also worth to ckeck it > user exits
I know that the permission is ok otherwise the Server would not start ... but maybe
we have to check the code again ....
thank's rj
This is also a funny thing i did not try to do an update ....
ANR3763E An error occurred while the catalog file used for client updates was downloading from
https://public.dhe.ibm.com/storage/tivoli-storage-management/catalog/client/catalog.json
ANR2547E A Centera device (4) reported error "Unknown
error (transid='server/3133732/WRITE_BLOB'),FP_SERVE-
R_ERR" during command BlobWrite. (SESSION: 10062)