Piskvor's questions

I have a box connected to a largish local IPv4 network (no IPv6 at all) through a single Ethernet interface; normally, it is assigned an IP address from the 172.16.0.0 range.

From time to time, however, somebody plugs in a DHCP server by mistake, and I get an address from the 192.168.0.0 range (and a bogus set of routes). Is there a way of telling dhcpcd (or another DHCP client for Linux) to only accept offers from a specific IP range? Or perhaps a way to filter DHCPOFFER at iptables level, dropping bogus offers?

DHCPCD's manpage lists blacklist and whitelist options, but these seem to apply to the server's IP address, which is in the correct/expected range.

I am only in control of this box (and a few others in a similar role); network management is beyond my control.

What I'm trying to do: provide an IMAP server as a facade for standardized e-mail access, without using on-disk mailboxes.

What I have:

• multiple TB of e-mails, stored in MySQL in a completely non-standard format
• a webservice providing access to these e-mails in a structured way (enumerate/create/delete folders, list/read/delete/send messages, etc.)

Currently, users access their e-mails through a web frontend or a smartphone app (both of which query the webservice for data, under the hood). This works; but I now need to provide access to the mails via IMAP as well.

I have been looking at various IMAP servers, and it seems that Dovecot allows for custom storage backends. However, all the examples I have found are using local mailboxes for storage.

What I'm trying to get to is an IMAP server which would query the webservice for folders/messages instead of looking into a local database/mailbox. (Is this even possible, or would I need to somehow fake the existence of a local file structure, e.g. by writing a FUSE module mimicking http://wiki2.dovecot.org/MailboxFormat/Maildir ?)

I have two 120 GB disks in a software raid1 array (i.e. mirrored), showing as /dev/md2. There's an ext3 filesystem on top of this, mounted at /.

# uname -a
Linux svnserv 2.6.26-2-amd64 #1 SMP Sun Jun 21 04:47:08 UTC 2009 x86_64 GNU/Linux

Recently, writes to the disk started failing with "write error: No space left on device". df showed that there's no more disk space. I removed some unused large files, approx. 5 GB, yet error persists.

df shows this:

# df -h
Filesystem            Size  Used Avail Use% Mounted on
/dev/md2              109G  104G     0 100% /
tmpfs                 471M     0  471M   0% /lib/init/rw
udev                   10M  100K   10M   1% /dev
tmpfs                 471M     0  471M   0% /dev/shm
/dev/md0               44M  9.8M   32M  24% /boot

Note that 104 out of 109 GB is used, yet there's 0 GB available.

I'm not running out of inodes:

# df -i
Filesystem            Inodes   IUsed   IFree IUse% Mounted on
/dev/md2             7208960   55273 7153687    1% /

mdadm seems to be happy (both disks are active and in sync):

# mdadm --detail /dev/md2
/dev/md2:
        Version : 00.90
  Creation Time : Mon Jul 20 18:53:01 2009
     Raid Level : raid1
     Array Size : 115218112 (109.88 GiB 117.98 GB)
  Used Dev Size : 115218112 (109.88 GiB 117.98 GB)
   Raid Devices : 2
  Total Devices : 2
Preferred Minor : 2
    Persistence : Superblock is persistent

    Update Time : Thu Jul  8 14:34:08 2010
          State : active
 Active Devices : 2
Working Devices : 2
 Failed Devices : 0
  Spare Devices : 0

           UUID : 1bb0fff8:38e78897:34693ac2:f3157c61
         Events : 0.71

    Number   Major   Minor   RaidDevice State
       0      22        3        0      active sync   /dev/hdc3
       1      22       67        1      active sync   /dev/hdd3

How can I make the unused(?) space available?