I use SAS EG 8.2.
And I know there is stop buttons on menu bar.
But sometimes when the requesting result of my proc sql program is little(??) bit large,
that stop buttons doesn't work.
It always goes where it hopes to go.
From time to time, it takes over 1 hour(. And now I'm waiting and asking for)...
So I hope to know how to terminate a running proc sql program at once.
There's a bunch of different possible issues here.
When you say proc sql, if you mean a job that is submitting code to a SQL database, you probably cannot force that job to quit. Your database administrator would have to do that, on the SQL database end - SAS usually cannot do anything about it.
If it's a SAS job only in PROC SQL, then often the stop button will be fine. But if not, and if it's not a local job, you may be able to use PROC IOMOPERATE to quit the job, depending on your permissions on the server. You may not be able to, as it's generally restricted to administrators, but again your SAS Admin may be able to. See Chris Hemedinger's post on the subject; he also developed an EG add-in that will allow that to work.
It is harder to control what SAS is doing when you use a front-end like Enterprise Guide to submit your SAS code instead of running SAS directly.
You could try disconnecting from the SAS Server and starting a new session.
Related
I'm having a bit of trouble with an old website that I have inherited from someone.
It throw errors regarding a connection pool being maxed out. When it happens the website would simply then just load the HTML and nothing from the database. When it is left for a while it would then work fine, or when I would recycle the IIS application pool in Plesk.
I have done a lot of reading and research but I can't quite work it out still.
The first thing I read was to look for was any code where the database connection was not closed after it had retrieved the information. I haven't found anything like this so far.
The next thing was I found a stored procedure sp_who2 which I was led to believe would give me the open connection but I'm a little confused as to if this is what I'm actually looking at.
When running sp_who2 I get the below.
Is this an open connection? Or is it simply my connection that is currently looking at the database through SQL Server Management Studio?
This database is currently on a shared hosting platform so I don't have the access needed to run some of the other commands that I found.
Ideally I will move the website off a shared host, but I'd like to find out the reason for this before I do. What I'm hoping to find is where the code needs to be adjusted to make it work.
Is this an open connection? Or is it simply my connection that is currently looking at the database through SQL Server Management Studio?
It's both. Your connection is coming from SSMS, and it is an open connection.
There are ways to get more detailed infomration than this. You can, for example, use sys.dm_exec_connections, along with other DMV's like sys.dm_exec_sessions, and construct a query which tells you a lot of information.
But rather than write a query like that yourself, I suggest you download Adam Machanic's excellent sp_whoisactive. It's just a stored procedure you create, which contains a query to pull lots of useful infomration from the system metadata, with the ability to provide options to customise the output. This procedure is, I would be confident to say, the "default" procedure that everyone eventually uses for this kind of query. I might write my own more limited query against the DMV's from time to time, but most of the time, sp_whoisactive can give you everything you want.
One of the parameters, for example, is #show_sleeping_spids. This will show you connections even if they're not actually running any query. Which is sort of funny given that the procedure name is "who is active", but the usefulness is clear. You would want to execute this procedure like so: exec sp_whoisactive #show_sleeping_spids = 1, and perhaps add other paramters besides. It has powerful filtering options too, but if you don't typically have a lot of connections, you probably can just grab the entire output and eyeball it for relevant info.
If you can't run this because of the permissions available to you by shared hosting, then your options are severely limited. If you are only getting your own session from sp_who2, but you know other stuff is running, then you don't have view server state permission. In such cases, sp_who2 only outputs your own session (because you are allowed to see information about your own session).
I'm in a database class and the teacher wants us to connect through ssh to an oracle database setup on a school server and it's been extremely frustrating. She wants us to turn in an sql file that will create all the necessary table, insert tuples, run certain select commands which I've found to be very hard to get an sql file with everything after i get everything right and I haven't found a way to test the sql file against the server and I don't think I have permission to drop tables anyway. Anyway my question is there a way I can take an sql file with create table and insert commands to convert it to something like an access .mba database or something local i can mess around with? and help would be greatly appreciated didn't find much help on google.
You seem to be confusing terminology a bit; SQL*Plus is a client application, and the database is a shared server resource. You want to create schema objects from an SQL file, I think. But anyway...
There's a very useful online resource for experimenting with bits of SQL in various flavours, SQL Fiddle. Technically not 'offline' of course, but I'm taking that to mean off your school's network, not necessarily completely isolated. You can create tables and run your inserts in the schema panel, and then run queries against that. Make sure you pick the right database product from the drop-down menu so you're using syntax that is valid for your class. You'll see a lot of answers here with links to demonstration fiddles.
That's great for a lot of things but if you want something a bit more robust and scalable, and entirely offline, you can install VirtualBox and get a pre-built developer VM image which gives you a ready-to-go Linux environment with a database installed and running. You can run whatever you want against that, you have SQL*Plus and SQL Developer available, and you can connect to the DB from your host machine if you want to. You can create and test your scripts against that, and in a format that will be closer to what you have to hand in than you'd use with SQL Fiddle.
This is much less work than installing the Oracle software yourself and learning how to create and manage the database, which I'm guessing is a bit more advanced than you need at the moment, based purely on the kinds of thing your question suggests you're dong at the moment. I think you'd learn a lot from the installation and build process, but I'd get comfortable with Oracle first, and maybe practice in a VM first as it's so much easier to trash it and start again when you mess something up.
If I wanted 'something local I can mess around with', I would go for a VM image. Mo posted a walkthrough of the VM setup as a comment to a previous similar answer, which you might find helpful.
"Something local I can mess around with" in terms of Oracle Database is Oracle Database 11g Express Edition. It's free and can be downloaded from oracle.com. You certainly can test sql-files run through sqlplus on Oracle Database XE.
To get the MS Access (GUI) feeling, download SQL Developer. It's free.
Best of luck!
Bjarte
Is there a view or internal sp to do this?
For example, I have a sp spGoesOnForSomeTime.
If I kicked this off then some individual closes my computer down, how can I see whether this is still running or not?
I realise I can use SQL Profiler and ActivityMonitor but I ultimately want to relay this information back through a web app.
EDIT: Apologies, it is not a local connection.
If it's run under your local connection, then it will stop and roll back, so you can be confident that it is not running.
If it's running under the context of another connection, you can use the sp_who stored procedure to see all of the activity (and active connections) on the server, and the cmd column should provide you with the command. If any of the records have your procedure name in their cmd column, then that will tell you that it's executing.
You may, however, want to take a more intentional approach and set a flag of some kind (a value in a row in another table, an extended property on the database or procedure, etc.) when the procedure starts, then reset it when the procedure finishes. This would also account for scenarios where your procedure gets called from within another procedure.
If running it from your local Management Studio, the connection will be broken (closed) when SSMS closes.
Any transactions will rollback, all locks will be released. If you're in the middle of some huge data changes, your proc stops running and the connection is still closed, but the rollback will continue anyway
So no need to check based on the facts given...
Depends whether you're running it using SqlServerAgent or just via Mgmt Studio.
As #gbn said, once you disconnect then any work using that connection will stop too.
I'd say create a scheduled Job and then it's independent of your connection.
If this is part of an application requirement then I would log the start and finish of all of the calls to this stored procedure then have a view against that to show which instances of the stored procedure are currently running.
If this is for maintenance of the database then I would use sp_who or sp_who2.
EDIT:
Actually, sp_who and sp_who2 will likely show you the currently running command, not necessarily the stored procedure name itself. You can try using the system dynamic view sys.dm_exec_sql_text, but that isn't always going to be accurate either.
I am looking at the activity monitor in MS SQL Management Studio 2005, and I see an entry from a login that I created, and it's using the tempdb, and the last command is always "SELECT INTO". And everytime I hit "Refresh" in activity monitor, that entry's "CPU" column goes up by like 60. The "program" of that activity is "MS SQL Management Studio."
I tried to kill that process, but it won't allow me, saying "cannot be killed". I disconnect from the database and reconnect, and that process is still there.
Is it gonna eventually crash my SQL Server because of continous activity? And is anyone familiar with this situation?
Thanks!
This is absolutely normal. The Select Into is selecting data into a temp table to show you what you see in the Activity Monitor. No need to panic. It will not crash your SQL Server.
Raj
That is you. More specifically SSMS. Typically it will SELECT INTO some temp table (something it calls, will do it actually) in order to get data from the server to display to you. Open another SSMS and you should see two of them.
Is there any way to map to a network drive by using a stored procedure? I have tried:
xp_cmdshell 'net use Q: [shared_network_drive] [pwd] /user:[username]'
but I got an error saying something like
'System error 1312 has occurred.'
'A specified logon session does not exist. It may already have been terminated.'
However, when I run the same command in a cmd console in the same box as the SQL server machine, there is no problem at all. It looks like that SQL process does pass correct logon credential information. Any way to do it by SQL SP?
In an other forum someone posted a tip about how to solve the System error 1312 has occurred. problem by specifying the username including its domain like this:
net use Q: \\[host]\[folder] [pwd] /user:[host]\[username]
I tried it and it worked for me.
I spent one hour, but the problem solves: You must specify [HOST]\[username]
If you write like this: u:/username - it doesn't work
Do you have a proxy account set up for xp_cmdshell? If you are not a member of sysadmin, it requires a proxy account. The documentation for xp_cmdshell and sp_xp_cmdshell_proxy_account will explain it in detail, but try executing the following:
EXEC sp_xp_cmdshell_proxy_account 'DOMAIN\Username', 'password'
Of course, replace the user and password information with the account information on your network under which you want this to execute.
WARNING: xp_cmdshell can be a HUGE hole in security. Make sure that at the very least the proxy account that you use has very limited security. Don't use any kind of admin account and don't use an actual user's account.
Why would you even want to do this?!? Even if it is possible, you should be carefully asking yourself why you would want to do so.
My suggestion: don't. My impression is that you need to change your approach--whatever purpose that network drive mapping might serve should be pursued in a more appropriate place outside of the database engine. Once you choose that appropriate place, actually invoking the mapping will probably become trivial.
Rod, you are right. It is a big risk to do that in SQL server and not all the SQL admins would allow to open this door. What I am doing is to migrate legacy process to a scheduled SQL job so that required data will be available for another department's application to fetch data in another scheduled process.
I guess that a windows service job might be a better application to pull data out and transfer the result to the required network drive.
Actually, use Win Scheduled Tasks tool is what I need. I can create a .Net app to do the job and then add the application to the scheduled tasks.
The problem happens when using Distributed File System, no matter which operating system you're using. Instead, get the actual name of a server hosting the DFS content.