our SC3 with sp2b on a winnt4 sp6 just stops accepting any logins.
Those already logged in still works . We are using floating users and there is at least 10 clients login left .When checking the log file
we only see the ip from the user trying to connect, When using the normal client.
what else can u check ??
//FzeroE
Can you post a part of your logfile here?
Did you change anything on your server just today?
(SP maybe?)
Is your key still ok?
Did you check the windows-nt - Eventview? Maybe there is something logged?
Frank
How many users are logged in?
Does it always stop at the same number?
For the users that are trying to connect, do they just get the splash screen when trying to connect and then a time-out?
------------Logfile opened------------
569 04/11/2002 10:14:17 Server session started on socket(236)
569 04/11/2002 10:14:17 Child scserver process 569 started on socket 236.
569 04/11/2002 10:14:17 ServiceCenter scserver pid (569)
569 04/11/2002 10:14:17 Release 3.0 SP2b (0004121517) System: 12670 (0x61e13c00)
569 04/11/2002 10:14:17 on PC running Windows NT (4.0) from xxx (140.x.x.x)
569 04/11/2002 10:14:17 connected to client at 140.x.x.x:3745
186 04/11/2002 10:14:54 Express connection accepted on socket(196)
186 04/11/2002 10:14:54 Express session created
------------Logfile opened------------
546 04/11/2002 10:14:54 Express session started on socket(196)
186 04/11/2002 10:15:40 Express connection accepted on socket(204)
186 04/11/2002 10:15:40 Express session created
These logins are unsuccesful.
And a working one
293 04/11/2002 08:54:16 Express session started on socket(204)
293 04/11/2002 08:54:16 Attaching to resources for system '12680' with key 0x61e13c00
293 04/11/2002 08:54:16 ServiceCenter scenter pid (293)
293 04/11/2002 08:54:16 Release 3.0 SP2b (0004121517) System: 12680 (0x61e13c00)
293 04/11/2002 08:54:16 on PC running Windows NT (4.0) from XXX (140.x.x.x)
293 04/11/2002 08:54:16 connected to express client at 140.150.43.81:1755
293 04/11/2002 08:54:16 Process will be monitored for a heartbeat
293 04/11/2002 08:54:20 SQL: Now using 229376 bytes for SQL buffers
293 04/11/2002 08:54:20 SQL: Now using 393216 bytes for SQL buffers
293 04/11/2002 08:54:20 SQL: Now using 557056 bytes for SQL buffers
293 04/11/2002 08:54:20 Issuing connect request to YYY
293 04/11/2002 08:54:20 SQL Message 5701 Level 0 State 2: Changed database context to 'sc'.
293 04/11/2002 08:54:20 Connection to SQL database 'YYY' as user 'sc' successful
293 04/11/2002 08:54:20 User AAA has logged in and is using a Floating license (18 out of a maximum 37)
293 04/11/2002 08:54:30 SQL: Now using 720896 bytes for SQL buffers
293 04/11/2002 08:54:30 SQL: Now using 884736 bytes for SQL buffers
127 04/11/2002 08:55:54 Europe/Middle 02/04/11 08:23:01 Problem 023P0041876 has been opened.:
186 04/11/2002 08:55:55 Express connection accepted on socket(196)
186 04/11/2002 08:55:55 Express session created
/FzeroE
Q U E S T I O N or T O P I C
Our ServiceCenter server started fine, including all of the normal background processes/agents. All users who are already logged into the system are working without problems,
and there are no abnormal error messages in the sc.log file. However, no other users can login to the system.
A N S W E R or C O N T E N T
Problem: The ServiceCenter operator file is missing, empty, or corrupt, or a ServiceCenter quiesce command was issued
Explanation: Users can only login to ServiceCenter if they have a valid operator record in the ServiceCenter operator file. This is true even if LDAP is used (although, with LDAP,
some operator information such as passwords may come from an LDAP source rather than directly from the ServiceCenter operator file). Occasionally, new ServiceCenter
adminstrators have been known to delete or reset the operator file accidentally. Also, it is possible (although highly unlikely) that the operator file can become corrupted and
therefore unreadable for some or all operator records. When any of these problems occurs, current users (i.e., already logged in), can continue their client sessions without
problems, but no new users can login and even existing logged-in users cannot start additional sessions.
Also, if an administrator has issued a quiesce command within ServiceCenter (MVS only), no new users can login until you shutdown and restart the ServiceCenter server.
Symptom:
1. When you start the ServiceCenter server, all of the normal background processes start. This means, for example, that on Windows NT systems the console message shows
the normal number of processes (typically 10-20) running.
2. The sc.log file contains no abnormal error messages.
3. Any users already logged into the system can operate normally.
4. Users who attempt to login all receive this error message: Invalid login name/password. Please try again. This occurs even with a login ID and password
used successfully in the previous session, or currently logged on in another session.
Corrective action:
1. If users are already logged into ServiceCenter, do NOT log them off until you complete your troubleshooting. This is especially true for any users logged in with administrator
privileges, e.g., SysAdmin.
2. If an administrator is currently logged into the system, use the administrator's session to access the Database Manager and search the operator file.
3. If the operator file is missing or empty, use the ServiceCenter Database Manager Unload and Load utilities to restore the operator file from a recent backup, or contact
Peregrine Customer Support for assistance.
4. If the operator file shows any possible signs of corruption, run the LFSCAN portion of the SCDBUTIL database utilities, then contact Peregrine Customer Support for repair
assistance.
5. If your ServiceCenter server runs on MVS and the operator file appears normal, check with your fellow ServiceCenter administrators to see whether anyone has issued a
quiesce command within ServiceCenter. You can also look for the string "quiesce" in the ServiceCenter job log. If the system has been quiesced, notify all logged on users,
then shutdown and restart ServiceCenter.
Maybe that helps you?
It says that the users get a error message...
Is that the case ?
Frank
hehe i forgot to remove one IP address :smile:
But we have some other rel error aswell like these
121 04/11/2002 08:54:15 0000: 52454C42 4C4B696E 63696465 6E747300 [RELBLKincidents.]
121 04/11/2002 08:54:15 0010: 00000000 00000000 00000000 00000000 [................]
121 04/11/2002 08:54:15 0020: 02000000 10000900 A0AF5F01 00000000 [.........._.....]
121 04/11/2002 08:54:15 0030: 00000000 08555704 C490973C 01842002 [.....UW....<.. .]
121 04/11/2002 08:54:15 0040: 12000000 00000000 00000000 01000000 [................]
121 04/11/2002 08:54:15 0050: E8571503 C0C35000 E04CF301 00000000 [.W....P..L......]
121 04/11/2002 08:54:15 0060: 00000000 00000000 00000000 00000000 [................]
121 04/11/2002 08:54:15 LINES 0070-0080 SAME AS ABOVE
These lines continues for some rows and ends with.
121 04/11/2002 08:54:15 Dump of Trigger String
121 04/11/2002 08:54:15 0000: 01747269 67676572 2E696E63 6964656E [.trigger.inciden]
121 04/11/2002 08:54:15 0010: 74732E62 65666F72 652E706F 73740002 [ts.before.post..]
121 04/11/2002 08:54:15 0020: 74726967 6765722E 696E6369 64656E74 [trigger.incident]
121 04/11/2002 08:54:15 0030: 732E6166 7465722E 75706461 74650003 [s.after.update..]
121 04/11/2002 08:54:15 0040: 74726967 6765722E 696E6369 64656E74 [trigger.incident]
121 04/11/2002 08:54:15 0050: 732E6265 666F7265 2E706F73 74000474 [s.before.post..t]
121 04/11/2002 08:54:15 0060: 72696767 65722E69 6E636964 656E7473 [rigger.incidents]
121 04/11/2002 08:54:15 0070: 2E616674 65722E75 70646174 650000 [.after.update.. ]
These errors comes every other minute BUT it is still possible to login .
But after this error msg the logins stops working .
525 04/11/2002 09:17:13 sm_alloc: Not enough shared memory available to allocate 328 bytes
525 04/11/2002 09:17:13
------ Shared Memory ------
525 04/11/2002 09:17:13 Current size 24000000
525 04/11/2002 09:17:13 Shared Memory Rel/ver 3.0
525 04/11/2002 09:17:13 Segmented space 19642728
525 04/11/2002 09:17:13 Big alloc space 4333056
525 04/11/2002 09:17:13 Unused 24216 (0%)
525 04/11/2002 09:17:13
------ By Type ------
525 04/11/2002 09:17:13 Type Allocations Frees Allocated
525 04/11/2002 09:17:13 ------------------- ---------- ----------- -----------
525 04/11/2002 09:17:13 Not named 3369 3293 9872
525 04/11/2002 09:17:13 User blocks 1564 1521 22016
525 04/11/2002 09:17:13 Messages 3762 3762 0
525 04/11/2002 09:17:13 Resource locks 42916056 42916019 1792
525 04/11/2002 09:17:13 Database Services 265 0 61440
525 04/11/2002 09:17:13 Cache overhead 2966 0 193760
525 04/11/2002 09:17:13 Application cache 68224 67488 1307088
525 04/11/2002 09:17:13 DBDICT cache 75160 74934 2439424
525 04/11/2002 09:17:13 Format cache 0 0 0
525 04/11/2002 09:17:13 SQL descriptor cache 434 400 58432
525 04/11/2002 09:17:13 Join/ERD/Type cache 296 0 29568
525 04/11/2002 09:17:13 **** not used **** 0 0 0
525 04/11/2002 09:17:13 Remote DBDICT cache 0 0 0
525 04/11/2002 09:17:13 Remote Record cache 0 0 0
525 04/11/2002 09:17:13 String Type cache 10434 10434 0
525 04/11/2002 09:17:13 IR Expert cache 8270374 8265565 4151456
525 04/11/2002 09:17:13 Publish/Subscribe 194127 193897 8784
525 04/11/2002 09:17:13 24x7 cache 1 0 512
525 04/11/2002 09:17:13 Web cache 0 0 0
As u kan see our shared memory is 24000000 and i think we tried to increase this value but then SC didn't start :smile:
this machine is a Dell 6300 with 4 Xeon's and 2 gigs of memory
/FzeroE
Do You use IR Expert heavily?
IR expert use alot of shared memory.
Thx for all posts and help sugestions
We are gonna try to restart it in a minute or so.
we are just waiting for the last users to finish their stuff.
but as dajfiel said
"For the users that are trying to connect, do they just get the splash screen when trying to connect and then a time-out?"
this is what happens .
Good luck.
No not to my knowledge . I don't think we even use IR . But we use Cognos for reports and we where able to retrive the operator file and then it should't be broken/corrupt or ??
we have restarted SC and everything works fine this time :smile:
a month ago managed we to crash both SC3 machine and the sql server !! when we had a similar problem way to go :smile:
Can the users search for tickets in free text?
If Yes you are using IR Expert and you should really consider increasing the shared memory.
Im no SC expert but there is option called "ir query" in sc next to basic and adv search
The message you see in the log file indicates that the "shared memory" setting in your sc.ini is not high enough.
Based on "rules of thumb" check if the current value is equal to (or better yet exceeds) the following:
5 MB + 0.6MB*no. of users
So in an environment with 50 users (including reporting clients, no IR Expert) this would amount to:
5+0.6*50=35MB => shared memory should be: 36700160
It could also be due to a sort of memory leakage. You can monitor the "state" of you shared memory by executing the following command:
scenter -reportshm
This will show you the amount of never used memory (Unused Space) and the currently not used memory (Free Space). By regularly checking the status (scheduled dump) of this report you might be able to identify trends - and act accordingly.
If it looks like available memory is only going down (Free Space) and the "release" doesn't seem to be aligned to the number of users logged in you will at least have to log a call with Peregrine Customer Support, and in the mean time reboot your server regularly (if the usage is liniar - you can determine the maximum up-time without running into memory trouble and reboot within that timeframe).
Regards,
Jos
you should add a entry in your sc.ini:
shared_memory:48000000
Greetings Frank.
Great !!!! i will check the memory tomorrow with scenter -reportshm and increase the shared memory later on
One other area to check is the following Registry setting (this is a definite issue with W2K with SC4, not sure about WINNT with SC3);
HKEY_LOCAL_MACHINE\System\CurrentControlSet\Contro l\Session Manager\SubSystems\Windows
Original Value
%SystemRoot%\system32\csrss.exe ObjectDirectory=\Windows SharedSection=1024,3072,512 Windows=On SubSystemType=Windows ServerDll=basesrv,1 ServerDll=winsrv:UserServerDllInitialization,3 ServerDll=winsrv:ConServerDllInitialization,2 ProfileControl=Off MaxRequestThreads=16
The value below allows over 300 users on the system in question, although results are dependent on machine specification.
Modified Value
%SystemRoot%\system32\csrss.exe ObjectDirectory=\Windows SharedSection=1024,3072,2048 Windows=On SubSystemType=Windows ServerDll=basesrv,1 ServerDll=winsrv:UserServerDllInitialization,3 ServerDll=winsrv:ConServerDllInitialization,2 ProfileControl=Off MaxRequestThreads=16
Having said all of this, the lack of shared memory needs to be resolved first.
How often does you server get a complete restart?
We only restart when we must or when changeing something or when it goes down :smile:
We are not resposible for the sc3 or the database
only the server itself .
Another country is makeing all the changes,they
also add functions to SC but it takes sevral hours for them to just begin look at our problem .
There are currently 1 users browsing this thread. (0 members and 1 guests)
Posting Permissions
Reply With Quote
Bookmarks