06-07-2004 02:09 AM - edited 03-10-2019 07:51 AM
This is a bit of a strange one. We are running ACS 3.2(1) on a Windows 2000 machine. We have about 30 groups setup for various users. The one group (Group 1) always tells us that we have 30 users more than are actually in the group. The group says 90 users but when you list the users there are only 60. I moved all the users to a new group and now it says there are 30 users in the group but when you list them, it gives you nothing. I have backed up the database, done a new installation of 3.2(2) on another machine and completed a restore to that box and I still get the same result. I'm trying to find out if the group is counting incorrectly or if there are 30 "phantom" users somewhere! I have only recently inherited the ACS boxes so I'm not sure when this problem started.
There don't appear to be any known bugs pertaining to this. Has anyone else seen this before?
Thanks!
Solved! Go to Solution.
06-17-2004 04:52 AM
We definitely fixed the issue and the case is now closed. What we did was I sent him a backup of the ACS server so he could look at it. He then sent me back a backup file saying they found the problem and to restore the backup file on ACS. The email the TAC guy sent me looked like this:
"Basically on the Appliance we cannot create a Dump.txt as we can do on ACS installed on Windows Server by the csutil -d option.
This dump.txt is a readable format of the database unlike the .dmp
I uploaded the .dmp sent by you on the ACS(Windows Server) at my end created a dump.txt, corrected it by running the perl script and then uploaded it back on to the ACS server by the csutil -l option. They I took a backup of the ACS and sent it to you. I check the same .dmp on the Appliance at my end to confirm the correction.
This basically an indexing issue, caused when the admin deletes users and its link pointer doesn't get removed in the registry causing the issue.
As discussed, regular backup and running the dbcompact should help preventing such issue.
I have attached the perl script, you can use it if required in future.
Hope this helps, feel free to contact me if you have more question. At this point I will go ahead and close the service request as discussed."
If you want I can send you the email of the script that the guy sent me. But evidently like he said and what I was thinking, it is a pointer issue in the database.
06-07-2004 06:34 AM
Hello,
Under this circumstance, I will perform the dbcompact that will clean up any dangling information in the database. Please, follow the procedure as described below, but before that please, BACKUP YOUR DATABASE COUPLE OF TIMES (in GUI, system configuration --> Backup) -
Thanks,
Mynul
06-14-2004 01:40 AM
Hi Mynul,
Thanks for your suggestion - I tried it and rebooted the box, but unfortunately no luck!
It seems we're stuck with it. Its not a major catastrophe but more of an irritation. We've moved the users to a new group and marked this group as unusable.
We're planning on upgrading to 3.2(2) soon so we'll build the box from scatch.
Thanks anyway!
06-15-2004 07:01 AM
I am actually having a very similar problem with 3.2(3) Build 11. I have TAC case 600241615 opened up on these random and weird occurances where it states I have 19 users when I only have 13. The even more weird thing to this is I had users who were telling me their 7920s were not working and looking for a password. I checked it out and it was asking for the LEAP password. I then checked the ACS and noticed the user was in the list and tried to click on the user to edit it. Much to my knowledge, the information entered was set to default (blank). I re-entered everything and clicked submit. I brought up the user list and it showed the user in the database twice. I checked out both users and the information was correct in both. So I deleted one of the users thinking I dont need both. The next day the same user came to me and said her phone was not working again. I looked at the ACS server and the same thing happened again. So I went through the same process and kept the second user because it seems if I delete one of them, it would happen again. This happened the next day with half of the other users. So now I have 6 users with duplicate userids in the database. I then run dbcompact on the console to where it showed me the right amount of users (13), however, the group portion of it still showed there are 19 users when I only see 13. To follow on what you said, it is definitely anuisance but we should figure out what this is so it can get resolved.
It is almost starting to look more like a databasing issue.
06-17-2004 12:39 AM
Please will you keep us updated as to the status of your TAC case?
Thanks!
06-17-2004 04:52 AM
We definitely fixed the issue and the case is now closed. What we did was I sent him a backup of the ACS server so he could look at it. He then sent me back a backup file saying they found the problem and to restore the backup file on ACS. The email the TAC guy sent me looked like this:
"Basically on the Appliance we cannot create a Dump.txt as we can do on ACS installed on Windows Server by the csutil -d option.
This dump.txt is a readable format of the database unlike the .dmp
I uploaded the .dmp sent by you on the ACS(Windows Server) at my end created a dump.txt, corrected it by running the perl script and then uploaded it back on to the ACS server by the csutil -l option. They I took a backup of the ACS and sent it to you. I check the same .dmp on the Appliance at my end to confirm the correction.
This basically an indexing issue, caused when the admin deletes users and its link pointer doesn't get removed in the registry causing the issue.
As discussed, regular backup and running the dbcompact should help preventing such issue.
I have attached the perl script, you can use it if required in future.
Hope this helps, feel free to contact me if you have more question. At this point I will go ahead and close the service request as discussed."
If you want I can send you the email of the script that the guy sent me. But evidently like he said and what I was thinking, it is a pointer issue in the database.
06-17-2004 04:59 AM
That's brilliant...
I would appreciate it if you could forward me that script!
Many thanks!
06-17-2004 05:09 AM
And, if you're interested, I just found the bug for this: CSCed19826.
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide