On Friday, April 28, 2006 @ 5:41 PM, Rich Duzenbury wrote:
Hi all,
I've got a looping mysqlmanager eating all the cpu on the system. On this server there are two instances of mysql. They get started just fine, they work fine, other than the cpu being eaten. For the time being, I've disabled mysqlmanager, and I've written a couple of shell scripts to fire up each instance, so things are ok.
I'm asking here, rather than the mysql lists because I'm pretty SuSE ignorant and I've just inherited this server.
My first thought was to upgrade to the latest mysql rpm's from mysql.com. In fact, there are other servers with Suse 10.0, running 5.0.18 and mysqlmanager that do not exhibit the looping behavior.
I went into yast, however, and I see little padlocks on all the mysql packages. Is that something done by SuSE, or by the previous administrator?
It is not done by SuSE, as far as I know. You can unlock those by simply right clicking on them and selecting "Keep" to remove the "Protected" status.
Here are the versions: MySQL-client-standard 5.0.13-0.sles9 MySQL-server-standard 5.0.13-0.sles9
I strace'd mysqlmanager, and it is looping as fast as it possibly can on a socket. Of course, I don't know what socket it's looping on, but I've got thousands and thousands of these: [pid 656] select(6, [3 5], NULL, NULL, {0, 0}) = 0 (Timeout) [pid 656] select(6, [3 5], NULL, NULL, {0, 0}) = 0 (Timeout) [pid 656] select(6, [3 5], NULL, NULL, {0, 0}) = 0 (Timeout) [pid 656] select(6, [3 5], NULL, NULL, {0, 0}) = 0 (Timeout) [pid 656] select(6, [3 5], NULL, NULL, {0, 0}) = 0 (Timeout) [pid 656] select(6, [3 5], NULL, NULL, {0, 0}) = 0 (Timeout) [pid 656] select(6, [3 5], NULL, NULL, {0, 0}) = 0 (Timeout) [pid 656] select(6, [3 5], NULL, NULL, {0, 0}) = 0 (Timeout) [pid 656] select(6, [3 5], NULL, NULL, {0, 0}) = 0 (Timeout) [pid 656] select(6, [3 5], NULL, NULL, {0, 0}) = 0 (Timeout) [pid 656] select(6, [3 5], NULL, NULL, {0, 0}) = 0 (Timeout) [pid 656] select(6, [3 5], NULL, NULL, {0, 0}) = 0 (Timeout) [pid 656] select(6, [3 5], NULL, NULL, {0, 0}) = 0 (Timeout) [pid 656] select(6, [3 5], NULL, NULL, {0, 0}) = 0 (Timeout) [pid 656] select(6, [3 5], NULL, NULL, {0, 0}) = 0 (Timeout) [pid 656] select(6, [3 5], NULL, NULL, {0, 0}) = 0 (Timeout) [pid 656] select(6, [3 5], NULL, NULL, {0, 0}) = 0 (Timeout) [pid 656] select(6, [3 5], NULL, NULL, {0, 0}) = 0 (Timeout) [pid 656] select(6, [3 5], NULL, NULL, {0, 0}) = 0 (Timeout) [pid 656] select(6, [3 5], NULL, NULL, {0, 0}) = 0 (Timeout) [pid 656] select(6, [3 5], NULL, NULL, {0, 0}) = 0 (Timeout) [pid 656] select(6, [3 5], NULL, NULL, {0, 0}) = 0 (Timeout)
I don't do Mysql (DB2 and Oracle, yes), but it looks like it's trying to get some resource and can't get it, so it keeps going back and trying again? Maybe the resource actually isn't available, or maybe the process can't get enough memory to handle the request, etc.
Since this server is only used to make backups (not replication) from another, I had the luxury of being able to wipe out all the data and restart the server. It still eats all the CPU.
I've been using debian distros for the past few years so I'm much more familiar with apt than yast. So, my question is - what is the proper way to proceed? Try to replace the binaries? With Yast? Without Yast? If I go with downloading and installing the rpms from mysql.com, what will happen to the yast database?
Thanks for any advice.
Regards,
Rich
Greg Wallace