pve-common.git
4 years agoadd bash completion handler
Dietmar Maurer [Wed, 2 Sep 2015 13:22:12 +0000 (15:22 +0200)]
add bash completion handler

4 years agoadd new property to support bash completions
Dietmar Maurer [Wed, 2 Sep 2015 11:31:17 +0000 (13:31 +0200)]
add new property to support bash completions

4 years agobump version to 4.0-18
Dietmar Maurer [Tue, 1 Sep 2015 15:47:47 +0000 (17:47 +0200)]
bump version to 4.0-18

4 years agoadded 'extra-args' CLI handling
Wolfgang Bumiller [Tue, 1 Sep 2015 09:32:30 +0000 (11:32 +0200)]
added 'extra-args' CLI handling

added 'extra-args' standard option
added 'extra-args' handling to PVE::JSONSchema::get_options
untainting 'extra-args' separately in RESTHandler::handle

4 years agobump version to 4.0-17
Dietmar Maurer [Fri, 28 Aug 2015 09:37:32 +0000 (11:37 +0200)]
bump version to 4.0-17

4 years agoAdded PVE::Network::tcp_ping to replace Net::Ping
Wolfgang Bumiller [Wed, 26 Aug 2015 08:38:27 +0000 (10:38 +0200)]
Added PVE::Network::tcp_ping to replace Net::Ping

We use Net::Ping twice in pve-storage (once for ISCSIPlugin
and once in GlusterfsPlugin, both with the 'tcp' variant.),
but Net::Ping doesn't support IPv6.

4 years agoTools::run_with_timeout improvement + hires alarm
Wolfgang Bumiller [Wed, 26 Aug 2015 08:38:26 +0000 (10:38 +0200)]
Tools::run_with_timeout improvement + hires alarm

The following situations could lead to the 'unknown error':
1) As commented, when the alarm triggered after the first
signal handler was installed and before the new alarm was
installed. In this case the $signalcount was increased,
and worse: the original signal handler was never called.

2) When $code died, since the call itself wasn't in an eval
block, we'd leave the eval block containing the inner alarm
signal handler. Then there's a time window from leaving the
signal block (and with that restoring the first installed
only-counting signal-handler) and reaching the code to
restore the previous alarm where the counting alarm handler
could get triggered by our own alarm set before running
$code. In this case at least the the old alarm would be
restored, but we'd still trigger the 'unknown error'.

The new code starts off by suspending the original alarm
before installing any signal handler, then installing the
timeout handler inside the first eval block. The $code is
then run inside another eval block to make sure we reach the
alarm(0) statement before restoring the old signal handler
and alarm timeout.

4 years agoAdd generic parse_host_and_port function
Wolfgang Bumiller [Tue, 25 Aug 2015 09:00:06 +0000 (11:00 +0200)]
Add generic parse_host_and_port function

Added a generic function to split a host+port string to the
host and port part supporting the two most common ipv6
notations beside domains and ipv4: with brackets for the
address or a dot as port separator.

4 years agoUpdate prerequisites for development environment
Wolfgang Link [Tue, 18 Aug 2015 09:06:27 +0000 (11:06 +0200)]
Update prerequisites for development environment

4 years agobump version to 4.0-16
Dietmar Maurer [Mon, 17 Aug 2015 07:43:57 +0000 (09:43 +0200)]
bump version to 4.0-16

4 years agofix bug #860: fix the upload from a VM when it is restricted.
Wolfgang [Fri, 14 Aug 2015 09:42:42 +0000 (11:42 +0200)]
fix bug #860: fix the upload from a VM when it is restricted.

4 years agobump cversion to 4.0-15
Dietmar Maurer [Mon, 10 Aug 2015 07:06:39 +0000 (09:06 +0200)]
bump cversion to 4.0-15

4 years agotap device name parse fix
Alen Grizonic [Fri, 7 Aug 2015 12:56:29 +0000 (14:56 +0200)]
tap device name parse fix

Signed-off-by: Alen Grizonic <a.grizonic@proxmox.com>
4 years agobump version to 4.0-14
Dietmar Maurer [Tue, 4 Aug 2015 06:59:56 +0000 (08:59 +0200)]
bump version to 4.0-14

4 years agoget_standard_option: do not overwrite defaults when they evaluate to false
Dietmar Maurer [Tue, 4 Aug 2015 06:58:09 +0000 (08:58 +0200)]
get_standard_option: do not overwrite defaults when they evaluate to false

4 years agofix a typo pointed out by Immo Wetzel
Wolfgang Bumiller [Mon, 3 Aug 2015 08:23:46 +0000 (10:23 +0200)]
fix a typo pointed out by Immo Wetzel

4 years agobump version to 4.0-13
Dietmar Maurer [Mon, 3 Aug 2015 14:13:55 +0000 (16:13 +0200)]
bump version to 4.0-13

4 years agofix tap_plug with vlan for non aware vlan bridges
Alexandre Derumier [Mon, 3 Aug 2015 13:07:37 +0000 (15:07 +0200)]
fix tap_plug with vlan for non aware vlan bridges

Signed-off-by: Alexandre Derumier <aderumier@odiso.com>
4 years agobump version to 4.0-12
Dietmar Maurer [Thu, 30 Jul 2015 07:00:17 +0000 (09:00 +0200)]
bump version to 4.0-12

4 years agoreplace /sbin/vconfig with /sbin/ip call
Wolfgang Bumiller [Wed, 29 Jul 2015 08:17:20 +0000 (10:17 +0200)]
replace /sbin/vconfig with /sbin/ip call

4 years agotap_plug : add support for vlan aware linux bridge
Alexandre Derumier [Fri, 24 Jul 2015 11:52:59 +0000 (13:52 +0200)]
tap_plug : add support for vlan aware linux bridge

Signed-off-by: Alexandre Derumier <aderumier@odiso.com>
4 years agobump version to 4.0-11
Dietmar Maurer [Mon, 27 Jul 2015 11:12:42 +0000 (13:12 +0200)]
bump version to 4.0-11

4 years agoadd ProcFSTools::read_proc_net_ipv6_route
Wolfgang Bumiller [Mon, 6 Jul 2015 08:03:38 +0000 (10:03 +0200)]
add ProcFSTools::read_proc_net_ipv6_route

4 years agoPVE::Network: add some common constants
Wolfgang Bumiller [Thu, 9 Jul 2015 06:47:08 +0000 (08:47 +0200)]
PVE::Network: add some common constants

added ipv4_reverse_mask and ipv4_mask_hash_localnet as they
are used in several places.

4 years agovalidate error message (typo) corrected
Alen Grizonic [Tue, 14 Jul 2015 09:00:57 +0000 (11:00 +0200)]
validate error message (typo) corrected

Signed-off-by: Alen Grizonic <a.grizonic@proxmox.com>
4 years agoimprove parse_config in JSONSchema and SectionConfig
Wolfgang Bumiller [Wed, 15 Jul 2015 08:09:41 +0000 (10:09 +0200)]
improve parse_config in JSONSchema and SectionConfig

The old code used string substitution for every line of the
input string, while perl can also iterate over all matches
via the /g re modifier, and can turn ^ and $ to act as
beginning/end of line via the /m modifier.

Effectively allowing a "match over all lines" via a simple
while ($data =~ /^.*$/gm);

The situation is a little different in SectionConfig because
there's a nested loop invovled which doesn't work with /g.
For this there are two options and I chose the safer one by
simply doing a split on newlines first.
The alternative would be to open the data as a
filehandle-to-string and use <$fh> to read lines, however
I'd have to throw in an eval{} to be sure to close the
handle afterwards.

5 years agoadded missing install build prerequisites for development environment
Thomas Lamprecht [Thu, 2 Jul 2015 12:17:41 +0000 (14:17 +0200)]
added missing install build prerequisites for development environment

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
5 years agobump version to 4.0-10
Dietmar Maurer [Fri, 26 Jun 2015 09:51:24 +0000 (11:51 +0200)]
bump version to 4.0-10

5 years agoadd veth_delete
Alexandre Derumier [Thu, 25 Jun 2015 05:06:42 +0000 (07:06 +0200)]
add veth_delete

Signed-off-by: Alexandre Derumier <aderumier@odiso.com>
5 years agoadd veth_create
Alexandre Derumier [Thu, 25 Jun 2015 05:06:40 +0000 (07:06 +0200)]
add veth_create

Signed-off-by: Alexandre Derumier <aderumier@odiso.com>
5 years agobump version to 4.0-9
Dietmar Maurer [Fri, 26 Jun 2015 05:56:28 +0000 (07:56 +0200)]
bump version to 4.0-9

5 years agorun regression tests before building the package
Dietmar Maurer [Fri, 26 Jun 2015 05:54:48 +0000 (07:54 +0200)]
run regression tests before building the package

5 years agotests: list/create/update network interfaces
Wolfgang Bumiller [Thu, 25 Jun 2015 09:54:34 +0000 (11:54 +0200)]
tests: list/create/update network interfaces

5 years agowrite_network_interfaces: improved sorting
Wolfgang Bumiller [Thu, 25 Jun 2015 09:54:33 +0000 (11:54 +0200)]
write_network_interfaces: improved sorting

This way option line positions between interfaces can be
preserved.

5 years agoPVE::INotify::write_network_interfaces: readability
Wolfgang Bumiller [Thu, 25 Jun 2015 09:54:32 +0000 (11:54 +0200)]
PVE::INotify::write_network_interfaces: readability

5 years agoimporting etc_network_interfaces tests
Wolfgang Bumiller [Thu, 25 Jun 2015 09:54:31 +0000 (11:54 +0200)]
importing etc_network_interfaces tests

5 years ago/etc/network/interfaces: deal with OVS allow- lines
Wolfgang Bumiller [Thu, 25 Jun 2015 09:54:30 +0000 (11:54 +0200)]
/etc/network/interfaces: deal with OVS allow- lines

* __read_etc_interfaces:
Delete OVS ports from "allow-$OVS_BRIDGE" option lines in
order to prevent them from being duplicated or kept after
removing the port from the bridge.

* __write_etc_interfaces:
Deleting unused OVSPorts has the side effect of them not
being written out at all. If, however, they are
physical interfaces they'll reappear the next time the
configuration is read, because they're added from
/proc/net/dev.
Fix: if the deleted interface matches the same condition as
in read_etc_interfaces, readd it with the standard options:
{exists => 1, method => manual }
This is a purely aesthetical change in order to make sure a
write()->read()->write() chain doesn't produce two
different files each write().

5 years agoread_etc_interfaces: missing iface initialization
Wolfgang Bumiller [Thu, 25 Jun 2015 09:54:29 +0000 (11:54 +0200)]
read_etc_interfaces: missing iface initialization

If no 'auto' line was read the interface hash has to be
initialized.

5 years agoPVE::INotify: testable interface parser
Wolfgang Bumiller [Thu, 25 Jun 2015 09:54:28 +0000 (11:54 +0200)]
PVE::INotify: testable interface parser

read_etc_network_interfaces now opens all /proc files and
passes them as filehandles to the actual implementation
which is now is __read_etc_network_interfaces.

Similarly write_etc_network_interfaces now prints the
content string returned by the implementation now named
__write_etc_network_interfaces.

This way a testsuite can use the __* variants directly to
test the parser, passing its own versions of the /proc
files.

5 years agowrite_etc_network_interfaces: updated header comment
Wolfgang Bumiller [Thu, 25 Jun 2015 09:54:27 +0000 (11:54 +0200)]
write_etc_network_interfaces: updated header comment

The header now mentions that the file is auto-generated and
should not be tempered with directly.
It now recommends to add custom network configurations via
source/source-directory lines.

5 years agoread_etc_network_interfaces: improved parsing
Wolfgang Bumiller [Thu, 25 Jun 2015 09:54:26 +0000 (11:54 +0200)]
read_etc_network_interfaces: improved parsing

* parsing ipv6 blocks
* parsing extra lines like source/source-directory/...
* merge multiple bridge_port lines into one
* write options only once

The returned config hash is not just the interface hash
anymore. Interfaces are now in its 'ifaces' member hash. All
unknown options (including mappings) end up in its 'options'
hash.

Added a comment describing the config hash's layout in
detail.

An interface can now have an ipv4 and an ipv6 entry, they
will be returned as a single interface with
address/netmask/gateway and address6/netmask6/gateway6
elements. Additionally a 'families' array is available
listing which families are available. Ideally we'll at some
point allow unhandled families to be kept too, however, now
that extra lines like 'source' and 'source-directory' are
preserved, it is recommended to move all custom
configuration into separate files to not interfere with our
interface parsing.

Options such as bridge ports or ovs_* will now be written
out only for the first interface. If multiple protocol
families of a bridge contain bridge_ports lines they will be
merged into the first interface.

5 years agoremoved write-only $gateway variable
Wolfgang Bumiller [Thu, 25 Jun 2015 09:54:25 +0000 (11:54 +0200)]
removed write-only $gateway variable

5 years agoread_etc_network_interfaces: less strict parsing
Wolfgang Bumiller [Thu, 25 Jun 2015 09:54:24 +0000 (11:54 +0200)]
read_etc_network_interfaces: less strict parsing

*) Don't require indented lines, instead know when to end a
   section.
*) Don't require empty lines between sections.
*) Fixed non-/greedy regex issue

And turned (.*)\s* into just (.*) as .* eats the whitespace
too.

5 years agobump version to 4.0-8
Dietmar Maurer [Fri, 19 Jun 2015 05:24:15 +0000 (07:24 +0200)]
bump version to 4.0-8

5 years agoPVE:Daemon start/restart with systemd
Alen Grizonic [Thu, 18 Jun 2015 17:00:12 +0000 (19:00 +0200)]
PVE:Daemon start/restart with systemd

Added some lines which make it possible to stop/start/restart daemon services with pve scripts,
avoiding the problem of systemd not recognizing the new status of the (re)started service processes.

The scripts now rely on systemctl.

Signed-off-by: Alen Grizonic <a.grizonic@proxmox.com>
5 years agobump version to 4.0-7
Dietmar Maurer [Tue, 9 Jun 2015 10:16:50 +0000 (12:16 +0200)]
bump version to 4.0-7

5 years agonew helper dump_journal to view systemd journal
Dietmar Maurer [Tue, 9 Jun 2015 10:15:41 +0000 (12:15 +0200)]
new helper dump_journal to view systemd journal

5 years agoremove superfluous substitution
Wolfgang Bumiller [Fri, 5 Jun 2015 09:23:11 +0000 (11:23 +0200)]
remove superfluous substitution

regexps are greedy so the first substitution eats up
multiple groups of double-colons up to the last one already.
So unless there's some intended hidden side effect this
statement has no obvious purpose.

5 years agobump version to 4.0-6
Dietmar Maurer [Mon, 1 Jun 2015 11:18:14 +0000 (13:18 +0200)]
bump version to 4.0-6

5 years agoadd trigger for pve-api-updates
Dietmar Maurer [Mon, 1 Jun 2015 11:17:30 +0000 (13:17 +0200)]
add trigger for pve-api-updates

5 years agobump version to 4.0-5
Dietmar Maurer [Thu, 28 May 2015 16:07:50 +0000 (18:07 +0200)]
bump version to 4.0-5

5 years agodefer some daemon setup routines
Wolfgang Bumiller [Thu, 28 May 2015 14:54:16 +0000 (16:54 +0200)]
defer some daemon setup routines

A first step towards untangling some of the intermingled data and
functionality setup tasks for the daemons:

Daemon::new now only validates and untaints arguments, but doesn't
perform any actions such as setuid/setgid until the new Daemon::setup
method which is now executed from Daemon::start right before entering
Daemon::$server_run.

CLIHandler::handle_cmd now takes an optional $preparefunc which is
called after handling 'printmanpod' and 'verifyapi'.

5 years agofix a regex typo in run_command
Wolfgang Bumiller [Thu, 28 May 2015 07:27:14 +0000 (09:27 +0200)]
fix a regex typo in run_command

m/|/ is always true as it effectively matches 'nothing or nothing
anywhere in a string'
looks like it was supposed to be m/\|/

5 years agobump version to 4.0-4
Dietmar Maurer [Wed, 27 May 2015 13:34:06 +0000 (15:34 +0200)]
bump version to 4.0-4

5 years agospice-proxy format now uses new address format
Wolfgang Bumiller [Wed, 27 May 2015 11:16:35 +0000 (13:16 +0200)]
spice-proxy format now uses new address format

spice-proxy used dns-name before which matches a combination of letters,
digits and dots, which happens to include ipv4 addresses but not ipv6
ones.
Since 'dns-name' sounds like it's only for names, I'm adding an address
format which is either a dns-name or an ip (including ipv6).

5 years agoprevent the use of AI_ADDRCONFIG
Wolfgang Bumiller [Wed, 27 May 2015 11:16:34 +0000 (13:16 +0200)]
prevent the use of AI_ADDRCONFIG

perl's IO::Socket::IP passes AI_ADDRCONFIG if no GetAddrInfoFlags are passed,
which is often useful but also causes it to error when explicitly trying to
bind to 127.0.0.1 when there are no _other_ IPv4 addresses present.

5 years agoupdate changelog
Dietmar Maurer [Tue, 12 May 2015 08:36:40 +0000 (10:36 +0200)]
update changelog

5 years agonew helper: getaddrinfo_all
Wolfgang Bumiller [Tue, 12 May 2015 07:51:12 +0000 (09:51 +0200)]
new helper: getaddrinfo_all

As it's commonly used in ipv6 support code a getaddrinfo wrapper passing
default flags and dealing with the (err,result) tuple was added.

5 years agoadd a socket family argument to next_*_port functions
Wolfgang Bumiller [Tue, 12 May 2015 07:51:11 +0000 (09:51 +0200)]
add a socket family argument to next_*_port functions

Instead of assuming a local address of 0.0.0.0, the next_*_port family
of functions now takes an optional packet family argument (AF_INET/AF_INET6),
used for ipv6 support.

5 years agobump version to 4.0-3
Dietmar Maurer [Fri, 8 May 2015 10:38:17 +0000 (12:38 +0200)]
bump version to 4.0-3

5 years agoadd utility to fetch the socket family for a hostname
Wolfgang Bumiller [Fri, 8 May 2015 10:24:44 +0000 (12:24 +0200)]
add utility to fetch the socket family for a hostname

5 years agooptional socket family parameter for create_reusable_socket
Wolfgang Bumiller [Fri, 8 May 2015 10:24:43 +0000 (12:24 +0200)]
optional socket family parameter for create_reusable_socket

5 years agoprovide Tools::unpack_sockaddr_in46
Wolfgang Bumiller [Fri, 8 May 2015 10:24:42 +0000 (12:24 +0200)]
provide Tools::unpack_sockaddr_in46

5 years agoUse IO::Socket::IP instead of INET
Wolfgang Bumiller [Fri, 8 May 2015 10:24:41 +0000 (12:24 +0200)]
Use IO::Socket::IP instead of INET

5 years agofix typo
Dietmar Maurer [Thu, 23 Apr 2015 07:14:48 +0000 (09:14 +0200)]
fix typo

5 years agoadd additional methods for IP address verification
Dietmar Maurer [Thu, 23 Apr 2015 06:19:03 +0000 (08:19 +0200)]
add additional methods for IP address verification

5 years agocleanup: remove unnecessary prefix
Dietmar Maurer [Wed, 22 Apr 2015 08:08:37 +0000 (10:08 +0200)]
cleanup: remove unnecessary prefix

5 years agobump version to 4.0-2
Dietmar Maurer [Wed, 22 Apr 2015 07:56:37 +0000 (09:56 +0200)]
bump version to 4.0-2

5 years agoregister new standard option pve-startup-order
Dietmar Maurer [Wed, 22 Apr 2015 07:49:15 +0000 (09:49 +0200)]
register new standard option pve-startup-order

We can use this for lxc and qemu.

5 years agoimprove error message
Dietmar Maurer [Fri, 3 Apr 2015 06:10:23 +0000 (08:10 +0200)]
improve error message

5 years agouse libspice-protocol-dev (instead of debian package)
Dietmar Maurer [Mon, 23 Mar 2015 11:40:03 +0000 (12:40 +0100)]
use libspice-protocol-dev (instead of debian package)

5 years agouse wiki style headers for README.dev
Dietmar Maurer [Wed, 4 Mar 2015 05:59:12 +0000 (06:59 +0100)]
use wiki style headers for README.dev

5 years agoupdate installation instructions
Dietmar Maurer [Mon, 2 Mar 2015 10:11:49 +0000 (11:11 +0100)]
update installation instructions

5 years agorename data to src
Dietmar Maurer [Fri, 27 Feb 2015 15:57:20 +0000 (16:57 +0100)]
rename data to src

5 years agobump version for Debian Jessie
Dietmar Maurer [Thu, 26 Feb 2015 10:19:59 +0000 (11:19 +0100)]
bump version for Debian Jessie

5 years agobump version to 3.0-24
Dietmar Maurer [Wed, 11 Feb 2015 05:38:44 +0000 (06:38 +0100)]
bump version to 3.0-24

5 years agonext_migrate_port: higher port range - all other ranges are 100 ports
Stefan Priebe [Mon, 2 Feb 2015 08:40:55 +0000 (09:40 +0100)]
next_migrate_port: higher port range - all other ranges are 100 ports

Signed-off-by: Stefan Priebe <s.priebe@profihost.ag>
5 years agoFix next_unused_port already bin detection
Stefan Priebe [Mon, 2 Feb 2015 08:40:54 +0000 (09:40 +0100)]
Fix next_unused_port already bin detection

unsecure migration does not use localhost to bind to. Instead it uses a specific IP on the node.
As we don't know that IP in this part of the code the only chance to check for a possible already
existing bind is to bind to all IPs.

Signed-off-by: Stefan Priebe <s.priebe@profihost.ag>
5 years agobump version to 3.0-23
Dietmar Maurer [Mon, 9 Feb 2015 06:02:29 +0000 (07:02 +0100)]
bump version to 3.0-23

5 years agoread_etc_network_interfaces : don't always add vmbr0
Alexandre Derumier [Sun, 8 Feb 2015 14:54:51 +0000 (15:54 +0100)]
read_etc_network_interfaces : don't always add vmbr0

Currently it's impossible to remove vmbr0 from network gui,

if we really want to remove it, or replace it with an OVS vmbr0

Signed-off-by: Alexandre Derumier <aderumier@odiso.com>
5 years agobump version to 3.0-22
Dietmar Maurer [Thu, 8 Jan 2015 13:49:59 +0000 (14:49 +0100)]
bump version to 3.0-22

5 years agoDaemon: change option name 'run_dir' to 'pidfile'
Dietmar Maurer [Fri, 2 Jan 2015 13:15:09 +0000 (14:15 +0100)]
Daemon: change option name 'run_dir' to 'pidfile'

5 years agoDaemon: new option to change user/group (setuid/setgid)
Dietmar Maurer [Fri, 2 Jan 2015 13:06:46 +0000 (14:06 +0100)]
Daemon: new option to change user/group (setuid/setgid)

5 years agoDaemon: make sure we kill all child processes on TERM
Dietmar Maurer [Fri, 2 Jan 2015 08:31:21 +0000 (09:31 +0100)]
Daemon: make sure we kill all child processes on TERM

5 years agoDaemon: add helper to create sockets
Dietmar Maurer [Thu, 1 Jan 2015 14:41:19 +0000 (15:41 +0100)]
Daemon: add helper to create sockets

Those sockets are not closed and reopened at restart.

5 years agoDaemon: catch finished worker earlier
Dietmar Maurer [Thu, 1 Jan 2015 13:59:04 +0000 (14:59 +0100)]
Daemon: catch finished worker earlier

Else we get zobies for 5 seconds at restart.

5 years agoDaemon: also terminate old_workers on stop
Dietmar Maurer [Thu, 1 Jan 2015 11:42:31 +0000 (12:42 +0100)]
Daemon: also terminate old_workers on stop

5 years agoDaemon: call terminate_server if we get unexpected errors
Dietmar Maurer [Thu, 1 Jan 2015 11:06:17 +0000 (12:06 +0100)]
Daemon: call terminate_server if we get unexpected errors

5 years agoDaemon: implement new option leave_children_open_on_reload
Dietmar Maurer [Thu, 1 Jan 2015 11:02:13 +0000 (12:02 +0100)]
Daemon: implement new option leave_children_open_on_reload

We want to be able restart daemons without closing existing connections.

5 years agoDaemon: correctly set got_hup_signal flag (before we call terminate)
Dietmar Maurer [Thu, 1 Jan 2015 09:57:10 +0000 (10:57 +0100)]
Daemon: correctly set got_hup_signal flag (before we call terminate)

5 years agoDaemon: raise error if we cannot parse lockfd
Dietmar Maurer [Thu, 1 Jan 2015 09:41:52 +0000 (10:41 +0100)]
Daemon: raise error if we cannot parse lockfd

5 years agoDaemon: remove useless parameter
Dietmar Maurer [Wed, 31 Dec 2014 16:40:04 +0000 (17:40 +0100)]
Daemon: remove useless parameter

5 years agoDaemon: untaint value read from ENV
Dietmar Maurer [Wed, 31 Dec 2014 16:28:33 +0000 (17:28 +0100)]
Daemon: untaint value read from ENV

5 years agoDaemon: log error and exit if something fails inside constructor
Dietmar Maurer [Wed, 31 Dec 2014 16:17:44 +0000 (17:17 +0100)]
Daemon: log error and exit if something fails inside constructor

5 years agoRESTHandler: define empty DESTROY to avoid useless warings
Dietmar Maurer [Wed, 31 Dec 2014 16:11:28 +0000 (17:11 +0100)]
RESTHandler: define empty DESTROY to avoid useless warings

We oserverd the following warning sometimes:

  (in cleanup) no such method 'main=HASH(0x38b2d48)::DESTROY'

Defining an empty DESTROY seems to avoid it.

5 years agoDaemon: do not wait for lock on restart
Dietmar Maurer [Wed, 31 Dec 2014 16:01:51 +0000 (17:01 +0100)]
Daemon: do not wait for lock on restart

5 years agoDaemon: cleanup logging
Dietmar Maurer [Wed, 31 Dec 2014 11:31:21 +0000 (12:31 +0100)]
Daemon: cleanup logging

5 years agoDaemon: keep lock file open during restart
Dietmar Maurer [Wed, 31 Dec 2014 10:54:27 +0000 (11:54 +0100)]
Daemon: keep lock file open during restart

So that nobody else can grab the lock

5 years agoDaemon: move ENV manipulation into new()
Dietmar Maurer [Wed, 31 Dec 2014 09:29:23 +0000 (10:29 +0100)]
Daemon: move ENV manipulation into new()

5 years agoafter_fork_cleanup: move code into separate sub
Dietmar Maurer [Wed, 31 Dec 2014 09:19:54 +0000 (10:19 +0100)]
after_fork_cleanup: move code into separate sub