thiesdiggity's questions

Does anyone know of a way to dynamically set the relay host when sending mail through Postfix? At what point does Postfix set the relay host? Is there some way that I can tell Postfix to use a relay host for a particular email (that is outside of global settings)?

Basically, when an email is to be sent to an outbound email address (Gmail, Outlook.com, etc.) I want to be able to choose which relay host to use based on which one has the best deliverability rating. And this could possibly change within seconds so it can't be a "global" postfix setting.

I've thought about creating a content filter which would lookup the "To", find the right relay host and then send the email using another application. However, that seems very awkward to do since we are already in Postfix.

Any help would be greatly appreciated!

Thanks.

I've been struggling with an issue that I am hoping someone can help with. Basically, I have setup a working Zen Load Balancer going from HTTPS to HTTP but I am struggling with getting it to work using HTTPS on both sides. When I configure HTTPS to HTTPS I get the following error messages in the syslog:

Jul 19 15:13:28 zenbal01 pound: (b730cb70) e500 can't read header
Jul 19 15:13:28 zenbal01 pound: (b730cb70) e500 response error read from xxx.xxx.xxx.xxx

The LB and the backend server has the same certificates installed so the SSL handshake should match.

Has anyone else gotten this to work or have any advice on what can be done to get the handshake to work?

Thanks!

I have a quick question. I need Zabbix to have access to CURL. However, I don't think that when it was installed the with-libcurl was set:

./configure --enable-server --enable-agent --with-mysql --enable-ipv6 --with-net-snmp --with-libcurl

Is it safe to run that command again even though we have Zabbix setup and monitoring all of our servers? We really need CURL support however we don't want to lose our current monitoring setup.

Thanks!

I am trying to setup HAProxy with MySQL failover with Ubuntu. I used a setup similar to this serverfault question, however I am getting the following error when starting haproxy:

[ALERT] 341/220001 (17405) : parsing [/etc/haproxy/haproxy.cfg:29] : unknown option 'mysql-check'.
[ALERT] 341/220001 (17405) : Error(s) found in configuration file : /etc/haproxy/haproxy.cfg
[ALERT] 341/220001 (17405) : Fatal errors found in configuration.

I even tried installing the lastest version of HAProxy (1.4.22). Does anyone know how to fix this? I have Google'd the heck out of it and can't find any solution.

Thanks for your time!

I have a problem that I can't seem to solve. We have a Ubuntu server setup with RAID-10 and two of the drives dropped out of the array. When I try to re-add them using the following command:

mdadm --manage --re-add /dev/md2 /dev/sdc1

I get the following error message:

mdadm: Cannot open /dev/sdc1: Device or resource busy

When I do a "cat /proc/mdstat" I get the following:

Personalities : [linear] [multipath] [raid0] [raid1] [raid6] [raid5] [raid4] [r$
md2 : active raid10 sdb1[0] sdd1[3]
    1953519872 blocks 64K chunks 2 near-copies [4/2] [U__U]

md1 : active raid1 sda2[0] sdc2[1]
    468853696 blocks [2/2] [UU]

md0 : active raid1 sda1[0] sdc1[1]
    19530688 blocks [2/2] [UU]

unused devices: <none>

When I run "/sbin/mdadm --detail /dev/md2" I get the following:

/dev/md2:
    Version : 00.90
Creation Time : Mon Sep  5 23:41:13 2011
   Raid Level : raid10
   Array Size : 1953519872 (1863.02 GiB 2000.40 GB)
Used Dev Size : 976759936 (931.51 GiB 1000.20 GB)
  Raid Devices : 4
  Total Devices : 2
Preferred Minor : 2
Persistence : Superblock is persistent

Update Time : Thu Oct 25 09:25:08 2012
      State : active, degraded
Active Devices : 2
Working Devices : 2
Failed Devices : 0
Spare Devices : 0

     Layout : near=2, far=1
Chunk Size : 64K

       UUID : c6d87d27:aeefcb2e:d4453e2e:0b7266cb
     Events : 0.6688691

Number   Major   Minor   RaidDevice State
   0       8       17        0      active sync   /dev/sdb1
   1       0        0        1      removed
   2       0        0        2      removed
   3       8       49        3      active sync   /dev/sdd1

Output of df -h is:

Filesystem            Size  Used Avail Use% Mounted on
/dev/md1              441G  2.0G  416G   1% /
none                   32G  236K   32G   1% /dev
tmpfs                  32G     0   32G   0% /dev/shm
none                   32G  112K   32G   1% /var/run
none                   32G     0   32G   0% /var/lock
none                   32G     0   32G   0% /lib/init/rw
tmpfs                  64G  215M   63G   1% /mnt/vmware
none                  441G  2.0G  416G   1% /var/lib/ureadahead/debugfs
/dev/mapper/RAID10VG-RAID10LV
                  1.8T  139G  1.6T   8% /mnt/RAID10

When I do a "fdisk -l" I can see all the drives needed for the RAID-10.

The RAID-10 is part of the /dev/mapper, could that be the reason why the device is coming back as busy? Anyone have any suggestions on what I can try to get the drives back into the array?

Any help would be greatly appreciated.

Thanks!

I have a quick question. How does one add emails to Postfix's queue without it delivering the emails? I am writing a script to email Postfix statistics (queue count, sent, etc) to our support staff. I want to test it when there are emails in Postfix's queue to see if its working correctly. However, every time I send an email from Postfix it gets delivered. :)

Thanks!

I have a quick question. I have two instances of Postfix running on an Ubuntu server and I want to write some scripts to email our server admin Postfix statistics (queue count, sent, etc). How would I get that information on the second instance?

When I run the following command:

postqueue -p

Is that the information of the first instance, second or both? If its not information on the second instance, how would I get that information?

Thanks for your help!

how do you get the Slave node of a DRBD pair to mount the data directory so data can be sync'd?

I went through the following guide: https://wiki.edubuntu.org/ClusterStack/Natty#Prepare_Partitions to get the 2-node cluster setup. I have the nodes connected but the data directory on the secondary node is never mounted so the data from node-1 is never replicated to node-2.

Pacemaker "should" be mounting the drive on the Node 2 with this command:

primitive res_fs ocf:heartbeat:Filesystem params device="/dev/drbd0" directory="/mnt/eoacluster" fstype="ext4" # Mounts the filesystem specifying the DRBD device and mount point.

Here is the output of drdb-overview:

Node 1:

0:eoa-ib  Connected Primary/Secondary UpToDate/UpToDate C r---- 
/mnt/eoacluster ext4 99G 188M 94G 1%

Node 2:

0:eoa-ib  Connected Secondary/Primary UpToDate/UpToDate C r----

Below is df -h

Node 1:

Filesystem            Size  Used Avail Use% Mounted on /dev/mapper/DATA-root
                      96G  4.3G   87G   5% /
none                  496M  232K  496M   1% /dev
tmpfs                 501M  3.6M  497M   1% /dev/shm
none                  501M  104K  501M   1% /var/run
none                  501M  4.0K  501M   1% /var/lock
none                  501M     0  501M   0% /lib/init/rw
/dev/sda1             228M   51M  165M  24% /boot
/dev/drbd0             99G  188M   94G   1% /mnt/eoacluster

Node 2:

Filesystem            Size  Used Avail Use% Mounted on /dev/mapper/DATA-root
                       96G  3.0G   88G   4% /
none                  496M  232K  496M   1% /dev
tmpfs                 501M  3.2M  498M   1% /dev/shm
none                  501M  104K  501M   1% /var/run
none                  501M  4.0K  501M   1% /var/lock
none                  501M     0  501M   0% /lib/init/rw
/dev/sda1             228M   51M  165M  24% /boot

/mnt/eoacluster directory does exist on Node 2.

What am I missing? Anyone have any thoughts/suggestions?

Thanks for your help!

I am having some problems setting a virtual IP and Linux (Ubuntu) cluster and I am hoping someone can help. This is the first node cluster I've tried to setup and I am not finding many resources in regards to this issue.

I am going through this guide to setup DRBD/Virtual IP on Ubuntu: https://wiki.edubuntu.org/ClusterStack/Natty#HA_Virtual_IP_Address

But I am having some major problems setting up the virtual IP. I've setup DRBD and it syncs and loads without any problems. But when I try to add the Virtual IP I get the following error message

pengine: [843]: info: native_color: Resource virtual_ip cannot run anywhere

Here is my CRM config:

node EOA-MAIL-IB-1
node EOA-MAIL-IB-2
primitive drbd_eoa ocf:linbit:drbd \
    params drbd_resource="eoa-ib" \
    op monitor interval="15s"
primitive fs_drbd ocf:heartbeat:Filesystem \
    params device="/dev/drbd0" directory="/mnt" fstype="ext4" \
    meta target-role="Stopped"
primitive virtual_ip ocf:heartbeat:IPaddr2 \
    params ip="192.168.1.254" nic="eth1" \
    op monitor interval="30s"
group res_group fs_drbd virtual_ip
ms ms_drbd drbd_eoa \
    meta master-max="1" master-node-max="1" clone-max="2" clone-node-max="1$
colocation mnt_on_master inf: res_group ms_drbd:Master
order mount_after_drbd inf: ms_drbd:promote res_group:start
property $id="cib-bootstrap-options" \
    dc-version="1.0.8-042548a451fce8400660f6031f4da6f0223dd5dd" \
    cluster-infrastructure="openais" \
    expected-quorum-votes="2" \
    stonith-enabled="false" \
    no-quorum-policy="ignore" \
    last-lrm-refresh="1336683121"

Here is the output of "crm resource list":

Master/Slave Set: ms_drbd
 Masters: [ EOA-MAIL-IB-1 ]
 Slaves: [ EOA-MAIL-IB-2 ]
Resource Group: res_group
 fs_drbd    (ocf::heartbeat:Filesystem) Stopped
 virtual_ip (ocf::heartbeat:IPaddr2) Stopped

I have two NICS installed on both servers.

Any help would be greatly appreciated.

Thanks!

I am in charge of setting up our company email infrastructure and have a question on how to setup redundant pop/IMAP servers. I understand that I can setup failover for inbound email using MX records but how do you setup the same type of thing for email retrieval? If our mail server goes down I don't want to have to go to every computer and change the email settings.

POP/IMAP uses A records for DNS? If so, would the solution be to manually change the A record to point to the backup mail server if the primary server went offline?

Any thoughts or suggestions would be greatly appreciated.

I am having an issue with one of my Postfix/Dovecot mail servers and I'm unsure how to fix the problem. I will try to explain it in detail, here it goes:

I have an Ubuntu server setup using Virtual hosting with Postfix, Dovecot and MySQL. We have one domain setup as a virtual domain, for this example I am going to use mail.example.com. Under that domain we have one email address. I have another server (MS Exchange) setup using another one of my sub-domains, ex.example.com.

The problem is that when I SMTP into the account on mail.example.com and try to send an email to an account on ex.example.com, I get the email returned back to us with an "unknown host" error.

Now, I know that the mail.example.com server can resolve the ex.example.com domain because I can ping/dig while SSH'd into it. I can also log into Postfix via Telnet and send an email to an ex.example.com mailbox.

I'm guessing that it has something to do with Postfix/Dovecot looking locally for the domain in the virtual domain list because of the tld domain (example.com)?

If that's the case, how do I get Postfix/Dovecot to only look locally for the entire URL (mail.example.com) and if it doesn't find it, send it to the correct server by looking up the MX/A records (which I know exist and are setup correctly)?

I have been working on this all day and any guidance would be GREATLY appreciated!

Thanks for your time!

I just configured 4 3TB drives in a RAID-10 configuration but the total disk space in the RAID is 4TB. I used the following guide to setup the RAID-10, RAID-10 Intall. Shouldn't the size if the RAID-10 array be 6TB? According to this RAID Calculator it should be. Anyone have any idea of what might have gone wrong in the install?

Thanks in advance for your time.

Ok, so I have read a bunch of material on-line and curious as to your thoughts on split vs monolithic disk in a production environment.

We are going to be building the VM server on a server that has 4TB of disk space in a RAID10 configuration. We have the OS (Ubuntu ext4) on a monolithic VM and are going to be adding another disk to the VM for the "data". Now since the new disk will have 3.5TB which is the best option, split disks or a monolithic disk? I have read pros and cons of each (Vmware Discussion) but I'm not sure what is the best option because of the disk size we going to be adding.

One other thought we had was to create multiple monolithic disks (1TB each) and add them to an LVM. Any benefit in this approach?

I have a quick question. How do I setup postfix to send an email to another server (Exchange Server) when sending to an email address that has a sub-domain of our main server. For example, say our main server is mail.example.com and we have a Exchange server setup to receive emails from exchange.example.com. We have the MX records setup in our DNS and it receives correctly if we send from a GMail account. However, when we try to send an email from a @example.com account we get the following error:

Host or domain name not found. Name service error for name=exchange.example.com type=A: Host not found

I believe Postfix checks for local mailboxes first and if its setup with the domain it delivers to the local account, but in this case the sub-domain accounts are located in another server. Anyone have any thoughts on what I need to do within Postfix so it doesn't look locally for the exchange.example.com mailboxes?

I found relay_domains directive within Postfix but that doesn't seem to fix it when I add the sub-domain.

Thanks for your help.

I have recently setup a DHCP-Server on an Ubuntu 10.04 server machine. The machine has two interfaces (eth0, eth1). I have the DHCP server setup correctly because computers connecting to the network get a DHCP IP address but they cannot connect to the outside world.

I believe it has to do with the two interfaces being mapped/bridged together. How do I properly map one interface to the other? Is it bridging or mapping?

I see on this page: https://help.ubuntu.com/community/dhcp3-server they do something like

mapping hotplug
    script grep
    map eth1

What does that exactly do?

In other places I see the following:

iface br0 inet static
bridge_ports eth0 eth1

Which is better or more efficient?

Sorry for all the questions, I'm a noob at setting up a DCHP Server.

Thanks for your help.

EDIT:

This question is in relation to a previous question I asked on ServerFault: Is it possible to setup a DHCP Server for localhost only?.

Basically I want to setup one interface (eth0) to connect to the outside world while (eth1) listens for DHCP requests from the VM's (bridged network setting in VMWare). It's setup on a server and the VM's are properly getting the DHCP IP's but they cannot see the outside world. :(

I used the configuration files from this post as a baseline to setup DHCP http://ubuntuforums.org/showthread.php?t=825294 (only difference we have 2 interfaces while they have 3) but unfortunately they didn't share their interfaces setup and its an old post.

I have a quick question. I have a bunch of VM's (VMWare Workstation) running on a Ubuntu server and have found that VMWare NAT (DHCP) service is unreliable and slow. I have to do use NAT instead of Bridging because I don't have enough static IP's for all the VMs.

Is it possible to setup the Linux Host (Ubuntu) to be a DHCP server but only for the local VM's?

The server has 2 NIC's so if I set eth0 to be a static IP, which connects to the outside work, and eth1 to listen for DHCP. Now, I am thinking if I don't want DHCP to broadcast I would just not connect a cable to eth1 and setup the VM's to connect to eth1 for the DHCP addresses. That way DHCP would not broadcast through my network but be listening on that interface.

Would that setup work?

EDIT

I have to use NAT or DHCP for the Guest VM's because the servers are in a data-center which does not offer DHCP and we don't have enough static IP's.

Thanks.

We have some servers in a data center but the data center does not offer a DHCP service. We want to have some computers setup with DHCP and others that have static IP's.

Currently our WAN cable going directly into a switch and statically setting the IP's on our computers. Now my question is if I connect the WAN cable to a router then connect the router to the switch, will the router act as a bridge/switch and still allow the network traffic to the computers with static IP's (note: some of the computers might have static IP's in different subnets)?

Is a router the only option to achieve this? I have read that I could possibly setup a DHCP service on one of the computers but will the other computers be able to "see" it for DHCP?

Thanks for your help.

EDIT:

Our setup is pretty complex because we use a ton of virtual machines. Right now we are using VMWare Workstation 7 (on 4 different servers) and because our servers are in a data center we have the VM's connected to the VMWare NAT interfaces (over 80+ VMs and do not want to statically set each with an IP). But VMWare's NATing performance is very slow and causing some of the VM's to intermittently lose their network connections.

To add to the complexity our data center has issued us to sets of IP stacks, which unfortunately are on different subnets.

So we want to move away from using VMWares NATing and setup a DHCP server or router.

Each of the VM Host servers have 2 1GB NIC card. Currently each has a static IP and I wanted to route half of the VM's to one (eth0) and the other half to the other (eth1) using VMWare bridging.

If I setup a DHCP server on each host (because I do not want to route all DHCP traffic (40+ VMs) through 1 nic - potentially a lot of data being sent), can 1 NIC provide the DHCP while the other connect to the internet? And if thats the case will the Guest VM's be able to get an IP address from the Host's DHCP service?

Each host is running Ubuntu and I found this link:

https://help.ubuntu.com/community/dhcp3-server (multiple interfaces section)

Will that configuration work for our setup?

Sorry, I realize this if complex and I truly appreciate help/feedback.

I hope someone can help me with this issue. We currently have VMware Workstation setup on an Ubuntu 10.04 host with NAT networking configured for the guests (b/c we don't have a router at the servers location). Now, every 10 hours or so on some of the guests the networking is dropped (not all of them, only a select few) for a couple of minutes. Has anyone else encountered this problem?

More details about the setup:

• 2 NAT networks setup (vmnet8 & vmnet9)
• 2 NIC's
• Half of the VMs on vmnet8 and the other half on vmnet9
• Running about 25 VM's
• 2 Quad Core CPU's with 40GB of memory (so processor and cpu should not be the issue)

Most of the drops occur on the vmnet9 network.

The strange thing about it is that we have another server (basically a mirror) that does not have the NAT dropping issue.

Anyone have a fix? I have been pulling my hair out trying to resolve this problem.

Thanks very much for your help!

Can anyone point in the right direction on how to remove SWAP out of a RAID1 configuration on Ubuntu 10.04 server? According to the Ubuntu Advanced installation guide, here, it said to add swap to the raid but I am having some serious performance issues. I would like to remove it from the array and recreate it, with the same size, on the two separate drives. Zero data loss on the other array is a must.

Its a remote computer so I would prefer to do it via command-line if possible.

Here is the output from cat /proc/mdstat

md0 : active raid1 sdb1[1] sda1[0]
19530688 blocks [2/2] [UU]

md1 : active raid1 sdb2[1] sda2[0]
957230016 blocks [2/2] [UU]

Swap is on the md0 device.

Thanks

I have a quick question about accessing a MySQL database via Postfix and adding a custom header to an email based on the result. So we have setup an Ubuntu server with Postfix which is using a MySQL DB for virtual addresses, now my question is this: Is it possible to lookup an address in the virtual addresses table and if it matches the search criteria, add a custom Header to the email? We would also need this to happen when the email arrives in case there are any forwarders attached to the email account.

I have found resources on how to add custom headers via Postfix: http://saaboke.com/?p=22, but I do not know how to incorporate that with MySQL. Any suggestions would be greatly appreciated.

Thanks for your time.