' or 'en'. On our switches (Dell S4148T-ON and
Dell N3248TE-ON running Dell OS10 and OS6 respectively) the interface
names begin with 'ethernet' and 'Gi'.
The commit adds those prefixes to the regex.
I've raised a Merge Request [1] which adds the additional prefixes to
the regex.
--
Regards
Jim
/config-dns-res.html
The MR includes an update to README.md with instructions on the use of
the new cluster option (--neutron-dnsmasq-dns-servers)
--
Regards
Jim Scadden (FloydianJim)
[1]
https://salsa.debian.org/openstack-team/debian/openstack-cluster-installer/-/merge_requests/48
a 'management' network so there's no reason to want/need to use
SSH over a ceph-cluster network.
--
Regards
Jim Scadden (FloydianJim)
[1]
https://salsa.debian.org/openstack-team/debian/openstack-cluster-installer/-/merge_requests/47
to have been introduced in commit 1b054ece [1] when the
list of cinder backends was changed from ['CEPH_1_perf1','CEPH_1_perf2']
to ['CEPH_1'].
I've raised a merge request [2] to amend the default backend from
'CEPH_1_perf1' to 'CEPH_1'.
--
Regards
Jim Scadden (FloydianJim)
[1]
https
this first
merge request in before spending too much time on this, to (a) try to
keep the commits as small as possible, and (b) find out as early as
possible if I'm heading in the wrong direction! I'm intending to add
firewall rules in the coming weeks.
--
Regards
Jim Scadden (FloydianJim)
[1]
https
also have an IP in the ceph-cluster network.
I've raised a merge request on salsa [1] to also allocate a ceph-cluster
IP to compute_is_cephosd nodes as well as cephmon nodes.
--
Regards
Jim Scadden (FloydianJim on OFTC)
[1]
https://salsa.debian.org/openstack-team/debian/openstack-cluster-installer
Package: openstack-cluster-installer
Version: 42.3.0~bpo12+1
Tags: patch
When the list of block devices is determined by enc.php there is
different code used for different node types. Some of these assume that
any drive which is not *da can be used as a non-OS block device. This
however is not
Package: openstack-cluster-installer
Version: 42.3.0~bpo12+1
Tags: patch
When using a dedicated ceph-cluster network, puppet fails to determine
a node's Ceph IP. This is due to the SQL SELECT query including the
following impossible combination:
networks.role='ceph-cluster' AND
Package: openstack-debian-images
Version: 1.73
Tags: patch
Due to the different naming convention for NVMe device nodes,
build-openstack-debian-image fails to set up a RAID array because it's
trying to use /dev/nvme0n11 instead of nvme0n1p1.
I've attached a patch which fixes this.
--
Regards
Package: openstack-cluster-installer
Version: 42.3.0~bpo12+1
Severity: minor
Tags: patch
During auto-racking the switchport name and NIC firmware version are
rejected by OCI due to containing the / character:
# lldpcli show n | grep -E '(SysDescr|PortID)'
SysDescr: Dell EMC Networking
Package: openstack-cluster-installer
Version: 42.3.0~bpo12+1
Tags: patch
When provisioning a new node the interface names in
/etc/network/interfaces are set to 'Array' if OCI is unable to find the
correct interface to use. This is due to
get_ethname_from_network_config() returning an array when
Package: openstack-cluster-installer
Version: 42.3.0~bpo12+1
Severity: minor
Tags: patch
When adding a new node to the OpenStack cluster there are a number of
PHP warnings which appear in the apache error log, which can potentially
be misleading when diagnosing problems.
I've attached a couple
Package: openstack-cluster-installer
Version: 42.3.0~bpo12+1
Severity: minor
Tags: patch
It looks like something went wrong with the patch which I provided in
#1053506, so the problem still exists. I've attached a rebased patch.
Original code:
if(sizeof($ret["data"]) == 0){
Current code (as at
Package: openstack-cluster-installer-agent
Version: 42.3.0~bpo12+1
The command used in openstack-cluster-installer-agent to determine the
speed of a NIC can in some cases return the string "null", but the script
only checks for an empty response when checking if no speed could be
determined
# ip
Package: openstack-cluster-installer-common
Version: 42.3.0~bpo12+1
The oci-system-serial command is currently unable to determine the
serial number on my libirtd/QEMU VMs as the fallback dmidecode commands
do not return any data. I've attached a patch to use the system-uuid
instead.
I also
Package: openstack-cluster-installer
Version: 42.3.0~bpo12+1
The puppetserver command for signing a client cert has changed in puppet
7. This causes errors when installing new nodes.
root@oci-01:~# puppetserver cert --help
puppetserver: 'cert' is not a puppetserver command. See 'puppetserver
Package: openstack-cluster-installer
Version: 42.3.0~bpo12+1
When PXE booting a new node, on its first report there is an error in
the apache logs:
[Tue Oct 03 08:48:00.052212 2023] [php:error] [pid 83221] [client
10.10.48.12:54344] PHP Fatal error: Uncaught mysqli_sql_exception: Unknown
Package: openstack-cluster-installer
Version: 42.3.0~bpo12+1
When testing out auto racking, the machine-guess-racking command returns
and error when there is no match, rather than reporting 'No racking info
could be guessed.' as expected
root@oci-01:~# ocicli machine-guess-racking
It looks like having brackets in comments in the config file breaks the
parsing of the config file in PHP. Updated patch attached.
--
Jim
diff --git a/etc/openstack-cluster-installer/openstack-cluster-installer.conf b/etc/openstack-cluster-installer/openstack-cluster-installer.conf
index
Package: openstack-cluster-installer
Version: 42.2.1~bpo11+1
Severity: wishlist
Tags: patch
Further to, and dependent upon, Bug#1028393, I would like to use
OpenStack Cluster Installer with packages from an internal aptly
mirror repository. Unfortunately as the packages hosted by aptly are
signed
Package: openstack-cluster-installer
Version: 42.2.1~bpo11+1
The BIOS versions on Intel servers exceed the 32 character limit of the
'bios_version' column in the 'machines' table ot he OCI database:
# for attrib in system-manufacturer system-product-name bios-version; do echo
-n "${attrib}: " ;
On Tue, Jan 10, 2023 at 03:50:15PM +, Jim Scadden wrote:
> I just noticed a minor typo in an error message produced when no value
> is given to the new --debootstrap-keyring-file parameter. Updated patch
> attached.
Trying again with the right file this time.
Cheers
Jim
diff --gi
I just noticed a minor typo in an error message produced when no value
is given to the new --debootstrap-keyring-file parameter. Updated patch
attached.
Cheers
Jim
.TH build\-openstack\-debian\-image 1
.SH NAME
build\-openstack\-debian\-image \- build a Debian image to be used with
OpenStack
.SH
Package: openstack-debian-images
Version: 1.72
Severity: wishlist
Tags: patch
I would like to use build-openstack-debian-image to install OpenStack
(a la OpenStack Cluster Installer) using packages from an internal aptly
mirror repository. Unfortunately as the packages hosted by aptly are
signed
Package: openstack-cluster-installer
Version: 42.2.1~bpo11+1
(Using osbpo.debian.net archive on bullseye)
When booting the live image built using
openstack-cluster-installer-build-live-image, agetty fails to autologin
and does not present any login prompt at all.
Checking the service status
tags 772318 + upstream patch
forwarded 772318 https://bugs.kde.org/show_bug.cgi?id=341640
thanks
I've raised this upstream on the KDE BTS with a patch. I've also attached a
copy of the patch here.
--
Jim Scadden
Description: Allow kmixremote to function correctly when using dash as /bin/sh
This has been fixed upstream in 4.14.3
http://commits.kde.org/kdelibs/38a89ca0195dedee30240647b86c7b6df6788723
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
forwarded 756492 https://bugs.kde.org/show_bug.cgi?id=328014
tags 756492 + upstream
thanks
--
Jim Scadden
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
nasty surprises.
REVIEW: 117044
FIXED-IN: 4.13.2
--
Jim Scadden
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
On Mon, Jun 09, 2014 at 11:43:46AM +0100, Jim Scadden wrote:
Attached is a proposed patch which I have forwarded upstream which
modifies the behaviour of the dialog box in question so that closing it
will return the user back to the previous dialog
After testing my patch for a while I do
/files/kdelibs-4.7.96-mimetypes.patch
There's also an upstream bug open for this :
https://bugs.kde.org/show_bug.cgi?id=303627
I've attached a copy of the Gentoo patch (updated against the copy of
kde.xml in kde4libs-4.13.1
--
Jim Scadden
--- a/mimetypes/kde.xml
+++ b/mimetypes/kde.xml
report it appears that
this was a problem with the package browser-plugin-libreoffice rather
than konqueror, so if the problem returns then a new bug can always be
raised against that package.
--
Jim Scadden
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject
/show_bug.cgi?id=330526
--
Jim Scadden
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
--
Jim Scadden
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
back to the previous dialog
--
Jim Scadden
Index: kde4libs-4.13.1/kio/kssl/sslui.cpp
===
--- kde4libs-4.13.1.orig/kio/kssl/sslui.cpp
+++ kde4libs-4.13.1/kio/kssl/sslui.cpp
@@ -75,66 +75,76 @@ bool KIO::SslUi::askIgnoreSslErrors
. This
has the side effect that if you eventually give up and accept the
certificate, there are numerous bad password dialogs which you have to
close.
--
Jim Scadden
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas
forwarded 750995 https://bugs.kde.org/show_bug.cgi?id=335994
tags 750995 + upstream patch
thanks
Bug forwarded to KDE BTS https://bugs.kde.org/show_bug.cgi?id=335994
with patches which provide a workaround
--
Jim Scadden
--- a/kimap/loginjob.cpp
+++ b/kimap/loginjob.cpp
@@ -535,7 +535,7
don't know; I don't have a Wheezy install handy.
I have just set up wheezy on a VM and cancelling the connection works as
expected. This bug only appears to affect jessie.
--
Jim Scadden
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe
in /var/log/auth.log ?
--
Jim Scadden
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
-10 kio/kfile/kdiskfreespaceinfo.cpp
http://commits.kde.org/kdelibs/9a8426ae2ed35ac900b89a623618fc158d24ee6a
--
Jim Scadden
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
gives 0, for the others authentication
failure 1.
This looks like the authentication is actually passing. Could you try
installing pamtester (which is availale for jessie/sid) and run the
following command and provide the results:
pamtester -v login username authenticate
--
Jim Scadden
.
At that point, the mail feeder will happily send user credentials over
to the untrusted server.
This is a problem with kio. Closing the dialog causes the default option
to be selected. I have raised this in KDE BTS along with a proposed
patch: https://bugs.kde.org/show_bug.cgi?id=335375
--
Jim Scadden
$?
Thanks
--
Jim Scadden
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
tags 631671 + upstream patch
forwarded 631671 http://bugs.kde.org/show_bug.cgi?id=335327
thanks
Patch forwarded upstream
--
Jim Scadden
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
requires the use of patch(es) in bug #631671 for kommander to run
--
Jim Scadden
Index: kdewebdev-4.12.3/kommander/executor/instance.cpp
===
--- kdewebdev-4.12.3.orig/kommander/executor/instance.cpp
+++ kdewebdev-4.12.3/kommander/executor
and debian/kommander.install
I have built packages using the patches on wheezy jessie and also used
pbuilder to build against sid.
Please let me know if any further modifications are necessary. If everything is
ok then I would be happy to submit the upstream patch to the KDE BTS.
Thanks
--
Jim
(No such file or directory)
access(/usr/lib/libkommanderwidgets.so, R_OK) = -1 ENOENT (No such
fileor directory)
The debian package contains .so.4.2.0 and .so.4 but not .so
The attached patch resolves the problem by including
libkommanderwidgets.so back in to the package.
--
Jim Scadden
--- a/debian
You reported a bug some time ago in kdm
Please could you confirm if this bug is still present or if it was
caused by a bug in the nVidia drivers as described in
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=651155
--
Jim Scadden
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ
--
Jim Scadden
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
49 matches
Mail list logo
