My apache/external ip/virtualmin all died after creating another domain but with a same administration username and password. For example:
.
Immediately upon clicking the "create server" button, I received a "page not found" error. No ping responses from my vps' IP. Multiple reboots did not solve the problem. The server was only accessible via an internal IP by SSH. It was usable there, but pointless if I was locked out from the internet.
Luckily, this was an initial install... so rather, I just reinstalled Centos 5 and Virtualmin.
Shouldn't Virtualmin detect whether there an existing user or not before trying to create a duplicate administration username?
Any clue what the heck happened, and how to fix this issue if this ever occurs again?
Howdy,
Hrm, the symptoms you're describing make it sound like something else is awry.
Even if for some reason the system were to create two users with the same name (which it's designed to test for before creating domains), that wouldn't cause your entire server to go offline and become unavailable.
It's difficult to point to what exactly the problem might be... sometimes having too little RAM can be troublesome, especially when there's no swap available. But, you mentioned rebooting, which would typically solve something like that.
If that were to happen again, I'd be curious what an "ifconfig" looks like (and having ifconfig output from a working system to compare it to would be excellent).
-Eric
So far it seems like it could be a memory thing. 512mb seems to not go very far. I have been working on optimizing everything as much as possible. Let's see what happens.
Howdy,
SpamAssassin and ClamAV are notorious for using up a bunch of RAM, so if you don't need/want both spam and virus scanning on your server, you might consider disabling one or both of them. Since both do serve an important role, another option is to run them both an a remote server if you happen to have another.
-Eric
Hi.
I just finished optimizing by following the guide: https://www.virtualmin.com/documentation/system/low-memory . And turned off mysql for my only virtual server (but left it on for virtualmin/webmin). Also, disabled all mail/spam/av modules.
For some my memory usage either has increased or stayed about the same as shown in "top -H" as 410MB of 524MB used , meanwhile the virtualmin cp page shows 220MB of 512MB used.
Oddly, I now have 10 mysqld processes at 22MB each and still imap/pop3/mailman processes, even after I have rebooted my server.
I was able to go and ultimately disable most of these services and on-boot via Webmin >> System >> Bootup and Shutdown.
The only thing left are the 10 mysqld processes... I have tried everything to drop it to one. To mention, there aren't any active websites either.
Memory usage currently: 289MB of 524MB (top -H) & 152MB of 512MB (webmin).
Yeah, shutting down processes like you described is an excellent way to free up some space.
As far as MySQL goes -- I believe there's one daemon for it, with multiple threads that can run. The individual threads are capable of showing up as processes when using "ps".
There are some things you can do to reduce memory usage in MySQL. I haven't tested any of these personally, so be certain to back up your current config before trying any of it... but if you google for "reduce mysql memory usage", a number of interesting tidbits come up.
These guys here show some sample my.cnf config files to use for a smaller memory footprint:
http://www.theodorenguyen-cao.com/2008/12/07/reduce-mysql-memory-usage/
Also, the MySQL Performance Blog has had a lot of interesting things in it over the years... you can see some posts regarding memory here:
http://www.mysqlperformanceblog.com/?s=memory
I hope that helps!
-Eric