Home / user-142812

Guillaume Fenollar's questions

Martin Hope
Guillaume Fenollar
Asked: 2016-12-30 02:09:04 +0800 CST
  • 0

On my CentOS 7, I use cat /proc/<pid>/limits to see the actuel limit set. I got confirmation that this method is totally accurate here because I'm using it on MySQL, which can't dump a specific database with >8000 tables in it. If, by using prlimit I set it to 10000, the dump finishes successfully.

Now, I'd like to avoid using prlimit, so I configured /etc/security/limits.conf file to match my needs, and if I run a bash with the according user (mysql), I can see the correct limit value configure when running ulimit -a.

The problem is that the process mysqld, ran by mysql user, is still set to a default value (here 1024), which leads to fail when dumping the DB. Only solution I have is to run prlimit on runtime but I'd like to undertand what is missing so my processes respect limits configuration at startup.

centos security process limits centos7
Martin Hope
Guillaume Fenollar
Asked: 2013-02-08 03:22:12 +0800 CST
  • 5

I'm facing some troubles right now with an OpenVPN server an former colleague set up (I'm not too into it ATM). We're using OpenVPN for a private network of ~200 clients (to be more specific, only 15 clients, with VMs on them, bridged to the same network, 192.168.1.0/24).

It worked nicely for years, but for some weeks, it's acting strange. I have a laptop in this network, which used to have IP 192.168.1.156. Here the interesting part of my OpenVPN server configuration:

ifconfig-pool-persist ipp.txt
ifconfig-pool 192.168.1.1 192.168.1.254 255.255.255.0

And in my ipp.txt file:

gfenollar.xxxxx.local,192.168.1.156

But when I restart my openvpn client, I got another IP, always different. Of course, it allocate me an IP already used by another OpenVPN client, which is insane! Here is the log of my client:

Wed Feb  6 12:36:29 2013 SENT CONTROL [xxx]: 'PUSH_REQUEST' (status=1)
Wed Feb  6 12:36:29 2013 PUSH: Received control message: 'PUSH_REPLY,dhcp-option DNS 192.168.1.5,ping 8,ping-restart 20,ifconfig 192.168.1.24 255.255.255.0'
Wed Feb  6 12:36:29 2013 OPTIONS IMPORT: timers and/or timeouts modified
Wed Feb  6 12:36:29 2013 OPTIONS IMPORT: --ifconfig/up options modified
Wed Feb  6 12:36:29 2013 OPTIONS IMPORT: --ip-win32 and/or --dhcp-option options modified
Wed Feb  6 12:36:29 2013 TUN/TAP device tap0 opened
Wed Feb  6 12:36:29 2013 TUN/TAP TX queue length set to 100
Wed Feb  6 12:36:29 2013 do_ifconfig, tt->ipv6=0, tt->did_ifconfig_ipv6_setup=0
Wed Feb  6 12:36:29 2013 /sbin/ifconfig tap0 192.168.1.24 netmask 255.255.255.0 mtu 1500 broadcast 192.168.1.255

So here I got 192.168.1.24 instead of .156. It happens to every of my hosts, but fortunately they have a private IP in interfaces file, in static, so they're kind of unallowed to modify the ip with "ifconfig" command, like it happens to me.

Can anybody point me out a way to investigate?

Thanks a lot in advance!

Guillaume

dhcp