Question The server or the system has no more resources . Try a larger -n

Hello Experts,

Good day!

We have an old legacy system 8.3b progress version (its outdated) no one supports it, recently we are facing issue while logging to the MFGPRO we see an error message i.e "The server or the system has no more resources . Try a larger -n", we checked the log files but do not see the error related to -n and also we see only connected users less than 200 , we have users 900 set in pf file.

While checking in promon we see the "1 16952 Auto TCP 2419 -83 5 14002" negative value for the current users...
Is this is bug?
When we got this error (in test) we saw the number of users connected to the server was 13, so this cannot be the -n issue...
OS is HP-UX B.11.00...
 

Rob Fitzpatrick

ProgressTalk.com Sponsor
"The server or the system has no more resources . Try a larger -n"
This is error 748. It is not necessarily caused by -n being too low. There are lots of KB articles that describe how to diagnose this situation.
Progress Customer Community
Progress Customer Community
Progress Customer Community
Progress Customer Community
Progress Customer Community

we checked the log files but do not see the error related to -n
Which log? The database log? Client log? Do you see error 748? Do you see any other errors?

we see only connected users less than 200 , we have users 900 set in pf file.
Which pf? What is its content? What is the command that launches the primary database broker? Where do you see the count of connected users? A screenshot would be helpful.

While checking in promon we see the "1 16952 Auto TCP 2419 -83 5 14002" negative value for the current users...
Is this is bug?
It is helpful to mention which promon screen you are referring to. I assume you mean R&D, 1, 17 (servers by broker). Yes, a negative number of connected users would be a bug.

Note that this is the "Chit Chat" forum. In future you would be better off posting it in a more relevant forum, like Database Admin, where it will likely be seen by more people who can help you.
 
Hi Rob,

Sorry, I will post the query in correct group, thank you for your reply...
We faced this issue in prod also last month, we did not get any clue form the db log, so we re-started the db...

Below is the db pf details... this is test db pf since we faced this issue today...
-B 1000
-bibufs 50
-c 250
-L 100000
-Ma 5
-Mi 2

-Mn 10
-n 300
-N TCP
-spin 10000
-bithold 800
-minport 14000
-maxport 18000

We do not see the error 748 in the db log file... we just see the below messages
06:09:10 SRV 1: Rejecting login -- too many users for this server. (1334)
06:09:11 SRV 1: Rejecting login -- too many users for this server. (1334)
06:09:12 SRV 1: Rejecting login -- too many users for this server. (1334)
06:09:13 SRV 1: Rejecting login -- too many users for this server. (1334)
06:09:14 SRV 1: Rejecting login -- too many users for this server. (1334)


-> We checked the no of users connected to db in promon tool, and with proshut dbname -C list | wc command (12 113 835)..
-> We the user count negative value in R&D 1 (Status Displays)->3(Servers)...
 

Rob Fitzpatrick

ProgressTalk.com Sponsor
Ok, now the picture is becoming clearer.

The -n 300 and -Mn 10 together determine the total number of database connections available: 312. The calculation is -n + -Mn + 2.

This is all connections: primary broker, servers (up to 10), secondary brokers, remote clients, self-service clients, utilities, daemons, etc.

However, with this configuration, the number of remote connections you can have is much more limited; it is -Mn (total servers) * -Ma (maximum clients per server). You have -Mn 10 -Ma 5, so you can have at most 50 concurrent remote clients. And that is only possible if the broker is able to spawn all 10 servers, using the port range 14000 - 18000. If ports in this range are already in use or are even defined in /etc/services, the broker will not use them. If enough of them are in use, this could limit how many servers can be spawned, and further reduce how many remote clients can connect.

Also, given that you are on a very ancient release of Progress, it is very likely that there are client/server bugs that are impacting you in some way.

You can check promon R&D 1,3 to see how many servers have been spawned and how many clients are connected to each one. If a server slot has not been used yet, its Type will be shown as "Inactive". Of course, if you are seeing negative numbers for Current Users on that screen, that information may not be useful to you.
 
Hi Rob,

Thanks for your help!

Below is the screenshot of the promon R&D details, when the issue occurred...

Sv Cur. Max. Port
No Pid Type Protocol Logins Users Users Num

0 15846 Login TCP 0 0 5 20000
1 16952 Auto TCP 2419 -83 5 14002
2 16972 Auto TCP 204 0 5 14004
3 18293 Auto TCP 414 0 5 14023
4 18326 Auto TCP 17 1 5 14025
5 18352 Auto TCP 3 0 5 14027
6 0 Inactive 0 0 0 0
7 0 Inactive 0 0 0 0
8 0 Inactive 0 0 0 0
9 0 Inactive 0 0 0 0
10 0 Inactive 0 0 0 0

Enter <return>, R, P, T, or X (? for help):
 
Hi Rob,

Thank you so much for your help....

We can see the there are inactive server available, so the issue could be the negative value for the server 1 shown ?...
I do not see any issue with the ports also and in db log file or the system logs I do not see any issue...
 
Top