[gw1500se]

Never mind, I found the problem. There was apparently a bad or corrupt receipt for that app on that particular machine. I removed it and now I can install via ARD with no problems.

[gw1500se]

I have a little more information that someone might be able to respond to. After a lot of hair pulling I decided to try running the package on the errant machine directly instead of via ARD (please don’t ask why I didn’t do this sooner 😳 ). It fails there as well telling me that there is nothing to install. How did it decide that since it is untrue and how can I fix it? TIA.

[gw1500se]

Maybe. Thanks for the reply. I did not know they had CLI interfaces. I’ll check it out.

[gw1500se]

Remember what happens when you ‘assume’. 🙂 I think what rankles me is that once you go outside of server admin to configure mail and have problems Premimum Care no longer covers you. While I can understand Apple not wanting to be a Unix support company, they really need to provide access to more postfix and spamassassin settings in the GUI. On the other hand, some of what they do provide generates configurations that are flat out wrong.

[gw1500se]

IMO the latter. I have a Linux machine and setting up and maintaining mail with ‘webmin’ works fine. Apple hides too much from the admin.

[gw1500se]

Maybe this will help:

[url]https://www.afp548.com/article.php?story=20040824063737872[/url]

[gw1500se]

The key diagnostic is the “Last data sent” message. Since it is “RCPT TO:”, it clearly means a connection has occurred and the ‘HELO’ handshaking was successful. Apparently your server is taking too long to respond to the “RCPT TO;” command. I would suggest you look at your postfix rules for something that would delay that response such as some kind of validation rule that is hanging or taking too long. Start by seeing which users time out and which do not (you seem to have implied that anyway) and finding some commonality.

Sorry I couldn’t be much more help but maybe fresh eyes will say something dumb enough to give you and idea. 🙂

[gw1500se]

I’m not sure about #1 as we don’t set mail quota here but we do set disk quotas. That works however, there is a bug in mail that does not respect the quota and hangs when the user exceeds it. At that point the users mail database is messed up and has to be re-imported.

On the other hand, #2 should work except when you change quotas the user must logoff before they take effect.

I hope that helps a little.

[gw1500se]

Its not very difficult but make sure the mail service is turned off while you do this. All you need to do is copy /var/spool/imap from the old drive to the new one (mailboxes for all users). The old drive is no longer needed. Make sure ‘cyrusimap’ is the owner when you are done copying (sudo chown -R cyrusimap:mail /var/spool/imap). Next, remove the mail database directory, if it exists (sudo rm -rf /var/mail) and make a new one (sudo mkdir /var/mail). Initiaize a new mail database (sudo /usr/bin/cyrus/tools/mkimap). You need to fix the owner for that as well (sudo chown -R cyrusimap:mail /var/mail). Now you are ready to rebuild the database from the copied mailboxes (sudo /usr/bin/cyrus/bin/reconstruct -i). Piece of cake.

Note that this assumes your database location is set to /var/mail/imap and the mail store location is set to /var/spool/imap using server admin. If you put them somewhere else you need to change the above accordingly.

[gw1500se]

I think I need to clarify what I am trying to do. Perhaps the title is a misnomer since it is not really promoting the replica per se that is the problem. My master OD server also has all the user home directories which are shares mounted at login on the local machine. Those home directories are also duplicated on the disaster site replica server. When I promote the replica to a master, I need to change the home directories of the users to mount the share from the newly created master. One would think that simply changing the home directory via Workgroup Manager would be sufficient. Apparently it is not. As I said, when the user tries to log on at the disaster site network, it still tries to mount the home directory from the no longer available original OD master.

[gw1500se]

Yes. Do you have your network routing definitions right on the server?

[gw1500se]

When you configure the VPN server one of the things you will need to do is set up network routing definitions. They tell the connected device what to send through the tunnel and what to send through the public network.

[gw1500se]

Never mind. I figured out where to put it in /etc/my.cnf instead. Thanks anyway.

[gw1500se]

It sounds like your install is messed up. The mysql.sock file is a communications socket generate when the server is started. You can’t really create one. When you downloaded MySQL did you get the source or the binaries?

Assuming you got the binary and it is version 5, be sure you also install the startup items package. Here are some things to check:

1) Edit /etc/hostconfig and make sure MYSQL=-NO-. That disables the Apple provided version. In addition there should be a MYSQLCOM=-YES-. That is added by the startup items package and is done to avoid conflict with Apple’s version.
2) The directory for everything is in /usr/local/mysql.
3) Edit /etc/profile and put /usr/local/mysql/bin at the beginning of the PATH. This assures you will be using the correct mysql commands. (Remember to run ‘. /etc/profile’ after you change it or close the terminal session and start a new one so the new path is active.)

[gw1500se]

Have you looked at ‘edquota’?

[Author]

Posts