I can't seem to figure out how to append to the default path in a supervisord program config.
I can reset the path:
environment=PATH="/home/site/environments/master/bin"
But when I try:
environment=PATH="/home/site/environments/master/bin:$PATH"
I see that supervisord doesn't evaluate $PATH.
Google wasn't a big help on this for some reason, I cannot believe I'm the first person to need this.
Supervisord must have support for this, any idea what it is?
I have a xen machine (DomU) hosting a few VMs.
One of the VMs is called router (10.0.0.1) and is the only one of the machines supposed to be reachable from the outside.
There are other VMs which are going to recieve traffic, but it has to be filtered.
Let's call one of these machines web (10.0.0.2), as it serves a webpage.
Packets coming to web follow this routing path:
xen DomU -> router -> web
DomU's configuration:
-A PREROUTING -d <external_ip> -p tcp -m tcp --dport 80 -j DNAT --to-destination 10.0.0.1:80
-A POSTROUTING -o xenbr0 -j MASQUERADE
router's configuration:
-A PREROUTING -d 10.0.0.1 -p tcp -m tcp --dport 80 -j DNAT --to-destination 10.0.0.2:80
-A POSTROUTING -o eth0 -j MASQUERADE
Everything works as expected.
Port 80 from the external IP is routed via router all the way to web.
There is a problem tho:
web sees 10.0.0.1 as the source for the connection, rather than the real client IP.
I guess this is caused by the fact that there are two DNATing iptables.
But why?
I have 3 Linksys SR2024 switches which are basically non-managed 24 port Gbit + 2 miniGBIC.
I now had to add another switch to the network, and my provider didn't have the SR2024 switch anymore, so I got the Cisco SLM2024 which was a bit more expensive. It's pretty much the same thing but with management (that I don't need).
So I've connected the SLM2024 to a SR2024 via Cat6 cable, and for some strange reason, I get no link.
If I connect any machine with a Gbit NIC to both switches, it links with 1Gbit autonegotiated. If I connect the SLM2024 to a non-Gbit switch (I have a cheap 4port ASUS switch), it will link just fine on 100Mbit full duplex.
Since the SLM2024 has management, I've tried to see if something is misconfigured on it's side, but it's not, it advertises 1Gbit and lower. (hence the machines connecting succesfully at 1Gbit).
Since the SR2024 that I'm trying to connect it to also connects successfully with another SR2024 and other Gbit machines, it means that it advertises Gbit too.
But for some reason when I link the SR2024 to the SLM2024 I get no link.
Please note that I've properly tested the wire.
Does anyone have any idea what's wrong?
I have two machines behind a firewall, with the ssh ports forwarded to 2201 and 2202.
When I
ssh host -p 2201
it asks if I trust the machine, I say yes, it gets added to ~/.ssh/known_hosts
Then I
ssh host -p 2202
It doesn't let me, because there's already an entry for this IP in ~/.ssh/known_host:1 (the file was empty when I started, so line 1 is the one added by the previous ssh run).
This happens on CentOS 5.4.
On other distros (I've tried Arch), it appears that ssh matches the known_hosts to the ports too, so I can have multiple fingerprints for multiple ports on the same host/ip without any problems.
How can I get this same behavior for CentOS?
I couldn't find anything in man ssh_config (or at least not without disabling fingerprint checking).
I've found a temporary solution. If the known_hosts file has more than one entry for the same IP, it will check all of them before concluding something is wrong. But I'd really hate this to be the solution.
I have a network behind a NAT with a few machines.
The machines are:
I also have mydomain.com.
server.mydomain.com is pointing to my connection's IP (single IP), which is the router, which forwards ports to server.
Server, has a httpd running, which serves different sites based on vhosts.
So I have site1.server.mydomain.com, site2..
The problem is that all the traffic is going thru the router, and when I check logs I always see the router's IP for everything (so it's hard to see who is running the script with the while(1)).
I would just ServerAlias site1.server.local, but most of the sites have a root URL saved somewhere on top of which other URLs are built, so I can't do that.
The solution for me would be telling dnsmasq somehow to answer to *.mydomain.com with server's IP.
Is this possible somehow?
I have a machine running Arch x86_64 2.6.30.
It's root is set on a raid5 array automatically mounted under /dev/md0.
This is done via the kernel params as follows:
kernel /vmlinuz26 md=0,/dev/sda3,/dev/sdb3,/dev/sdc3 md=2,/dev/sda2,/dev/sdb2,/dev/sdc2 rootfstype=ext4 root=/dev/md0 ro
This used to work fine, however sometimes, it will fail to assemble the md0 array.
What happens is that everything goes normal, the kernel probes for block IDs, it finds matches, then waits for the array to assemble 10 seconds.
This is usually instantly done, however sometimes it waits 10 seconds, after which it timeouts and drops to a recovery console which I can't use because it doesn't accept any input. I think this may be because the only keyboards I have available are USB keyboards (even tho the keyboard works in the GRUB menu).
When this happens, I just have to reboot and the array will mount just fine.
Btw, this happens around 30% of the time.
It happens after clean shutdowns.
It can happen more than once in a row.
Since it fails to mount the rootfs, it can't write in any logs.
Has anyone ever seen something like this?
Any ideas why it might be happening?