| Title: | Mailworks-unix |
| Notice: | V2.0.4 now available -- see Note 4.37 5 |
| Moderator: | TAMARA::NEUMAN::Neumann |
| Created: | Wed Jun 02 1993 |
| Last Modified: | Tue Jun 03 1997 |
| Last Successful Update: | Fri Jun 06 1997 |
| Number of topics: | 1384 |
| Total number of notes: | 5851 |
Hello, We have a customer that reported a problem with Mailworks V2.0.2 consuming a high amount of swap space ( X488grecv processes). The customer investigated a following scenario that causes the reported problem. A TeamLinks user X sets an autoforwarding e-mail address in his Mail profile (in TeamLinks V2.5) equal to the e-mail address of TeamLinks user Y and vice versa. Such scenario triggers oscillating of one message and causing Mailworks server is consuming a high amount of swap space. This continues up to 100% of swap and crash of the process. The solution is to stop x488 processes and deleting the message from MTA queue, but our customer is unsatisfied with such solution. So, is it possible to detect or prevent to oscillating of such messages or is it a way to disable autoforwarding in user Mail profile? The configuration of our customer is MAILworks V2.0.2, MB400 MTA V2.0, Digital UNIX V3.2D, DECnet/OSI V3.2B. MAILworks server and MTA are on the same machine. The other problem which customer indicates is the very long delay (cca 3-10minutes) on Server side in case that Teamlinks user attaches attachement bigger than 2.5MB. There is no network problems and it is indicated more than 80% server resources consuming by MailWorks process in manipulation with this message. I'd be happy if someone would help me with solving these problems. I'll appreciate any information. Regards Petr Sindelar, Digital Prague, CZ [Posted by WWW Notes gateway]
| T.R | Title | User | Personal Name | Date | Lines |
|---|---|---|---|---|---|
| 1361.1 | Simple solution | TAMARA::lamac.zko.dec.com::Neumann | Stan Neumann | Mon Apr 07 1997 17:50 | 26 |
Sorry to take so long to answer - You don't specify how much swap space was allocated, and how large the mss and mcs processes got. However, I would assume that the customer ran out of swap space because the processes grew too large for the swap space. I'm afraid that I don't understand why the customer does not accept the solution. Clearly it does not make sense for address x to forward to address y and vice versa. That configuration is going to create all kinds of problems for both the servers and the network. Running out of swap space is only one of the problems you will see. Clearly, the solution is to turn off the circular forwarding. On the performance problem - can you have the customer run the script suppscr.sh which is located in the directory /usr/opt/DMW/unsupported. (This should be run from the root account), and send me the output. I would like to see how the machine is configured. You can send the file to me at [email protected]. Thanks, Stan | |||||
| 1361.2 | Problem Cause Specification | NETRIX::"[email protected]" | Petr Sindelar | Thu Apr 10 1997 04:15 | 30 |
Hello, Stan thanks for your responce I have to specify reason for circular forwarding (causes described problems): TeamLinks User A sets autoforward function to User B. User B dont know that it happens and sets his autoforward to User A. This same scenario is around more than 2 Users. At the End this functionality causes system collaps (swap is consuming up to 100% and in case that is there no system administrator, who detect this status, Messaging system going down). On other site this thing, that system automatically could not detect this status, could represent service refusal attack at system, becouse this status could causes single user by autoforward to himself. I think that it is very important for our (maybe for other too) to fix this problem. Stan, I will send you output file with configuration of system. Thanks for responce. I'll appreciate any information. Regards Petr Sindelar, Digital Prague, CZ [Posted by WWW Notes gateway] | |||||