Public bug reported:
Jammy installer failed.
ProblemType: Bug
DistroRelease: Ubuntu 22.04
Package: ubiquity 22.04.15 [modified:
lib/partman/automatically_partition/question]
ProcVersionSignature: Ubuntu 5.15.0-25.25-generic 5.15.30
Uname: Linux 5.15.0-25-generic x86_64
NonfreeKernelModules: zfs
Public bug reported:
[83212.756483] kscreenlocker_g[81781]: segfault at 44 ip 7f1091c48fe5 sp
7ffcbfebb860 error 6 in libQt5Core.so.5.15.2[7f10919ef000+30e000]
[83212.756500] Code: 28 00 00 00 0f 85 bd 05 00 00 0f b6 44 24 43 48 81 c4 88
00 00 00 5b 5d 41 5c 41 5d 41 5e 41 5f c3 0f 1f
It looks to be 'an interesting mystery' we're chasing. This system is
in production, so the results below are with the whole 'snooping engine'
off as without it the whole thing dies. As such, I don't think the
contents of the fdb and mdb tables mean much. The setups below are
unchanged, they
P.S. The reason this is a security issue is-- there is now an address on
the host that the guest also 'knows' and it sits on the bridge giving
access to all the other guests on the bridge. Most admins will not
'just know' they need rules to block fe80 traffic generated by host
interfaces--
I need to repeat: in sysctl.d put this line in a file, then reboot,
then your test setup will show the failure:
net.ipv6.conf.all.autoconf = 0
Otherwise, in your test setup the tables are populated, then you delete
the addresses, but the L3/4 code engaged by even a little time with the
fe80:...
** Changed in: linux (Ubuntu)
Status: Incomplete => Confirmed
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to the bug report.
https://bugs.launchpad.net/bugs/1959702
Title:
Regression: ip6 ndp broken, host bridge doesn't add vlan
Public bug reported:
After a normal day's operations, the monitors switch off after an hour
or so. Every morning since Impish, upon waking up and unlocking the
system: 2 of the monitors wake, the rest remain in sleep mode. The
only recourse is to do a graceful poweroff and cold boot. There is
On 2/5/22 18:59, Jay Vosburgh wrote:
> Harry,
>
> I am attempting to reproduce the behavior you describe, but have been
> unable to do so. Could you clarify some of the configuration specifics,
> as follows:
>
> Starting with step 2,
>
> "2. On the host, create a bridge and vlan with two ports,
On 2/5/22 18:59, Jay Vosburgh wrote:
> Harry,
>
> I am attempting to reproduce the behavior you describe, but have been
> unable to do so. Could you clarify some of the configuration specifics,
> as follows:
>
> Starting with step 2,
>
> "2. On the host, create a bridge and vlan with two ports,
Still exists 2/2022. KScreen. After a sleep, it's random which monitors
are detected at all. After a sleep some, but not all the monitors come
back at all, and the ones that do are horizontally aligned, the stored
alignment is lost. The only way to fix it I've found is to reboot.
--
You
Yup, those failures were to do with an old radeon chipset on an ancient
server.
On 2/1/22 17:33, Seth Arnold wrote:
> Sounds good, thanks:
>
> [0.00] Linux version 5.11.0-49-generic (buildd@lcy02-amd64-054)
> (gcc (Ubuntu 10.3.0-1ubuntu1) 10.3.0, GNU ld (GNU Binutils for Ubuntu)
>
Public bug reported:
I have a multi-display setup, a few traditional HD monitors.
Under 21.04, every time I choose 'Take a new screenshot' with the rectangular
region area capture mode-- the spectacle just disappears as if it was never
launched. The system resumes as if Spectacle was never
Public bug reported:
System call lsetfilecon in the guest fails, it worked when the host ran
Groovy.
How to repeat/test virtiofs lsetfilecon regression that breaks dpkg,
prevents upgrades/fixes.
In the guest, for example fedora:
[root@registry1 ~]# getenforce
Permissive
[root@registry1 ~]# cat
** Also affects: apparmor (Ubuntu)
Importance: Undecided
Status: New
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1894902
Title:
2 apparmor items missing for 'groovy'
To manage
S5000PSL - usb creator fails to boot using groovy daily build 9/12,
9/13. Works normally with unetbootin. Daily build did boot perhaps a
week or 10 days ago. (Installer failed & crashed when 'something else/
grub bios boot only .. no efi .. but that's another issue. 'Something
else' did work
This is still an issue 9/13/20, ubuntu mate groovy daily build. pre-
uefi server won't boot w/ubuntu usb creator, works with unetbootin.
Whatever the change was, it happened within the last 10 days or so. An
earlier daily build did boot on that same machine without trouble.
--
You received
server was bare-metal S5000PSL
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1883040
Title:
groovy daily won't boot anymore on some older BIOS boxes
To manage notifications about this bug go to:
Public bug reported:
need to add
~# cat /etc/apparmor.d/local/usr.sbin.ntpd
/run/systemd/userdb/io.systemd.Machine wr,
/etc/ssl/openssl.cnf r,
otherwise you get
Sep 8 16:07:44 noc4 kernel: [ 1843.949193] audit: type=1400
audit(1599599264.140:52): apparmor="DENIED" operation="connect"
Public bug reported:
Attempting to install Groovy using cd image. Normal until bootloader
install. The system is not EFI capable (older server, some VMs).
Installer advises without EFI the system will crash, etc. Though there
is a grub/bios partition, and a btrfs install proceeded normally up
Otto, thanks for the repeated reminder. Kindly note upstream includes a
ubuntu repo that fixes this problem. Interested people can get the
answer to this problem by following this link:
https://downloads.mariadb.org/mariadb/repositories/#distro=Ubuntu_release=focal
Thanks Otto.
It would be an appreciated add-on to the bug reporting system if your
firm would add a little alert when it accepts a bug submission to a
package that's not in the 'main' repository so we users get a gentle
reminder we ought not expect help of the sort the 'main' packages get.
--
That's it? A conclusion with no reasons, no exploration of options?
How about a check in the upgrade routine that makes a list of currently
installed packages with no reverse dependencies, that do not appear in
the new release repos, and warn the user if they consent to the upgrade
those
Public bug reported:
The clear, simple pop-up graphic advising of a new long term release can
lead to major upset with Ubuntu in a way that can be avoided, as
follows:
Just for the purpose of this wish: Let's call an installed package that
provides a capability of interest to the user that's
I agree. No issues here these many days.
On 7/6/20 9:42 AM, Łukasz Zemczak wrote:
> Ok, I think it's as good as it gets, waiting any longer won't make much
> difference. Thank you for the verification - releasing!
>
--
You received this bug notification because you are a member of Ubuntu
Bugs,
Public bug reported:
Mariadb 10.3 as provided by Ubuntu and shipped in LTS is compiled
against YaSSL version 2.4.4, which supports a maximum tls version of 1.1
as I understand it. See:
https://ubuntuforums.org/showthread.php?t=2420831
Focal minimum tls requirement is higher, tls v1.2 as
Still happening in clean focal install, mate desktop.
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1605224
Title:
at-spi-registryd.desktop not responding on shutdown
To manage notifications about
Christian,
Since to test whether the fix I provided was complete you'd have to
install freeipa-server to see all the bloat from apparmor, then try the
fix and tweak it that would be enough given the sample error message.
But, since you asked, here's a zmore /var/log/syslog.7.gz | grep appa >
Public bug reported:
Setting up python3-cephfs (15.2.1-0ubuntu2) ...
/usr/lib/python3/dist-packages/ceph_volume_client.py:358: SyntaxWarning: "is
not" with a literal. Did you mean "!="?
group_id = group_id if group_id is not 'None' else None
Public bug reported:
/usr/share/hplip/base/utils.py:2063: SyntaxWarning: "is" with a literal. Did
you mean "=="?
if weburl is "" or weburl is None:
/usr/share/hplip/check-plugin.py:116: SyntaxWarning: "is" with a literal. Did
you mean "=="?
if log_level is 'debug':
Seth,
Thanks for the note! I've made the change.
Harry Coin
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1884980
Title:
patch so apparmor complain->enforcing
To manage notifications ab
Public bug reported:
Kindly notice https://bugs.launchpad.net/ubuntu/+source/krb5/+bug/1885024
which won't be fixed in kerberos since freeipa changes the log destination. So
freeipa needs to add a systemd drop in to allow the logging.
krb5-kdc.service drop in:
[Service]
ReadWriteDirectories=
Public bug reported:
Jun 24 11:29:34 registry1 krb5kdc[1244]: Couldn't open log file
/var/log/krb5kdc.log: Read-only file system
can be fixed by adding
ReadWriteDirectories= /var/log
to the krb5-kdc.service unit
** Affects: krb5 (Ubuntu)
Importance: Undecided
Status: New
--
Public bug reported:
Because /var/log/syslog gets bloated with sssd apparmor related
messages, I put the following in /etc/apparmor.d/local/usr.sbin.sssd
then I changed sssd from 'complain' to 'enforcing' mode. I put this on
a heavy sssd vm running freeipa server that also is running the gui
Public bug reported:
Setting up python3-mysql.connector (8.0.15-2build1) ...
/usr/lib/python3/dist-packages/mysqlx/expr.py:972: SyntaxWarning: "is" with a
literal. Did you mean "=="?
return "({0})".format(dimension[0]) if len(dimension) is 1 else \
No issues so far. Thanks for checking.
On May 20, 2020 5:30:40 PM CDT, Dan Streetman <1873...@bugs.launchpad.net>
wrote:
>@hcoin I assume no more segfaults with the updated pkg? can we mark
>this verified for eoan?
>
>--
>You received this bug notification because you are subscribed to the
With the line not commented, upon each and every startup in all cases
one sees this:
May 19 11:37:25 email1 systemd[1]: Starting SSSD NSS Service responder socket.
May 19 11:37:25 email1 sssd_check_socket_activated_responders[72216]: (Tue May
19 11:37:12:251510 2020) [sssd] [main] (0x0010):
Public bug reported:
Notice
ipa-client-install
creates /etc/sssd/sssd.conf
but changes in the sssd process's socket approach calls for that file to change
/etc/sssd.conf from
...
[sssd]
services = nss, pam, ssh, sud
...
to
[sssd]
#services = nss, pam, ssh, sud
otherwise the sssd service either
python3-yubico specifically, not python-yubico
** Package changed: ubuntu => python-yubico (Ubuntu)
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1878994
Title:
python3-yubico/focal: "is" with a
Public bug reported:
Setting up python3-yubico (1.3.3-0.3) ...
/usr/lib/python3/dist-packages/yubico/yubikey_config.py:478: SyntaxWarning:
"is" with a literal. Did you mean "=="?
if slot is 1:
/usr/lib/python3/dist-packages/yubico/yubikey_config.py:483: SyntaxWarning:
"is" with a literal. Did
While dropping python-netaddr may remove a 2.0 issue, notice the bug
report was for python3-netaddr. The 3 bit is important and doesn't fix
this bug.
On 5/15/20 3:12 PM, Kai Kasurinen wrote:
> python-netaddr (0.7.19-4) unstable; urgency=medium
>
>* Drop python2 support; Closes: #937942
>
Public bug reported:
Setting up python3-netaddr (0.7.19-3) ...
/usr/lib/python3/dist-packages/netaddr/strategy/__init__.py:189: SyntaxWarning:
"is not" with a literal. Did you mean "!="?
if word_sep is not '':
'is' and == are not the same, tests will fail that succeeded on some x86
cpython
Segfaults attracted to me? That's possible. Or, I'm the only ubuntu
user that actually reports bugs. It's lonely out here.
More seriously, I think there's a problem with the apport system more
generally. What's supposed to happen after a crash is a window pops
up upon log-in, offering
crash report.
I suppose there ought to be another crash report/bug report about how this
crash report didn't get auto-uploaded upon detection as well.
** Attachment added: "_usr_sbin_qemu-ga.0.crash"
Public bug reported:
Completely new/fresh focal / mate install. Crash during initial apt install of
many packages after first log-in, including qemu-guest-agent. No other
important load other than apt running. OS running in a vm as a guest
(obviously...) No idea what caused it. Including
Good detail here on why this can go horribly wrong if ignored:
https://adamj.eu/tech/2020/01/21/why-does-python-3-8-syntaxwarning-for-is-literal/
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1878939
Public bug reported:
During apt install from an otherwise blank focal/mate:
setting up python3-cephfs (15.2.1-0ubuntu2) ...
/usr/lib/python3/dist-packages/ceph_volume_client.py:358: SyntaxWarning: "is
not" with a literal. Did you mean "!="?
group_id = group_id if group_id is not 'None' else
Trying it now on 1 of 4 previously affected systems. Accepted all
'proposed' items, including changes to libvirt. Will advise.
On 5/14/20 3:19 PM, Brian Murray wrote:
> Hello Harry, or anyone else affected,
>
> Accepted systemd into eoan-proposed. The package will build now and be
> available
Still on eoan. Staying with Ubuntu depends on freeipa support.
Presently only the freeipa client is supported past eoan, so the future
is unclear. Can the fix be backported?
On 5/13/20 3:03 PM, Dan Streetman wrote:
> @hcoin are you still on Eoan or have you moved up to Focal yet? If
> you're
And again:
root@noc1:~#
Broadcast message from systemd-journ...@noc1.1.quietfountain.com (Tue
2020-04-28 12:32:15 CDT):
systemd[1]: Caught , dumped core as pid 25084.
Broadcast message from systemd-journ...@noc1.1.quietfountain.com (Tue
2020-04-28 12:32:15 CDT):
systemd[1]: Freezing
Just tried it three times, all normal, nothing in the logs except a few
netfilter reports of correctly dropped packets.
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1873607
Title:
systemd segv
Public bug reported:
Proximate syslog:
Apr 18 16:26:56 gate2 rssmonitor.py[2001]: settings: Running: ip -all netns del
Apr 18 16:26:56 gate2 systemd[1]: /lib/systemd/system/dbus.socket:4:
ListenStream= references a path below legacy directory /var/run/, updating
crash log
** Attachment added: "_usr_lib_systemd_systemd.0.crash"
https://bugs.launchpad.net/ubuntu/+source/systemd/+bug/1873607/+attachment/5356610/+files/_usr_lib_systemd_systemd.0.crash
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to
This has been happening perhaps 3 times every 7 days, no obvious cause I
can see. The system is a virtual machine that's just a router. No
user activity, nobody logged in to the mate gui. Wish I had more to
offer by way of a clue but that's it.
--
You received this bug notification because
Yes, the upstream patches appear to have resolved this.
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1854129
Title:
regression: recent eoan patch killed ceph osd pool create
To manage
Yes, the upstream patches appear to have resolved this.
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1868157
Title:
Regression: ceph dashboard upgrade crashes totally w/SSL
To manage
Public bug reported:
apt install php-imagick fails complaining about overwriting /etc/php.
This is the work-around:
dpkg -i --force-overwrite ./php-php-imagick_3.4.3~rc2-2ubuntu7_amd64.deb
dpkg: error: cannot access archive
'./php-php-imagick_3.4.3~rc2-2ubuntu7_amd64.deb': No such file or
This looks to be relevant, upstream:
https://tracker.ceph.com/issues/38378
** Bug watch added: tracker.ceph.com/issues #38378
http://tracker.ceph.com/issues/38378
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
James, As you'll note by the upgrade log below, your thought above that
the dashboard was not changed needs a second look. Notice 2020-03-17
16:54:39, the dashboard module was in fact changed, and it broke SSL.
Does Canonical have a working ceph cluster anywhere that uses ssl on the
dashboard?
Just vanilla eoan on a vm with only the ceph subsystem loaded. No osds
on that vm. Are you saying you have a working dashboard that uses ssl?
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1868157
Public bug reported:
After the latest upgrade to ceph, the previously normal and well working
dashboard crashes on load. There were no configuration changes, just
the update. The dashboard is protected by an SSL cert that is valid.
The only change was the upgrade. Was this tested by anyone
Thanks Dan!
On 2/13/20 1:24 PM, Dan Hill wrote:
> pr#28869 fixed upstream in 14.2.3. This version is in the eoan-updates
> and focal pockets.
>
> For UCA, both train-proposed and ussuri-proposed have this fixed, but these
> have not yet landed:
> ceph | 14.2.2-0ubuntu3~cloud0 | train
Steps to reproduce:
Install the package.
Run the program.
Notice it fails either silently or with a default obscure error message.
Browse the program which is a shell script.
Notice it runs several programs as sub programs.
Notice those programs are not dependencies of the package.
See the
Public bug reported:
Notice the script that produces the backup relies on pv, socat, sockstat
and netcat.openbsd among other packages. These are not dependencies of
mariadb-backup, but should be. Without them, this package aborts
without any indication as to why.
Also, mariadb-server has a SST
The affected .deb file is mariadb-backup n.b.
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1858673
Title:
package missing dependencies
To manage notifications about this bug go to:
Timo,
You might take a look at /etc/sssd/sssd.conf
Consider changing
services = ifp
#services = nss, pam, ifp, ssh, sudo
The reason is that unless you change this, systemctl is-system-running
reports degraded instead of running, with messages akin to
Dec 9 17:59:25 registry1
Using the ppa, the upgrade to the primary server was successful. Then the
replica install was successful, other than, at the end:
...
Restarting named
Updating DNS system records
DNS query for registry1.1.quietfountain.com. 1 failed: All nameservers failed
to answer the query
Good to know. I was using ubuntu eoan.
On 11/27/19 11:18 AM, Timo Aaltonen wrote:
> for the record, ipa-replica-install works fine on the debian vm's that I
> have set up for this (and finally had a go at replicating 4.8)
>
> my goal is to eventually have it all tested with a CI system
On all the monitors, mgrs and mds's I gave the command:
find / -xdev -path "*/__pycache__/*.pyc" -delete
Then rebooted them sequentially.
Upon giving the same command to create a pool,
ceph --verbose osd pool create qfblockdevsnoc2 32
the same as above was emitted, but then this happened:
pool
The lock up happens also using the dashboard web interface. You can see
the server side monitor gets the command then either loses it or waits
forever somehow.
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
Public bug reported:
After applying recent eaon normal upgrades to an otherwise vanilla
system (4 osd hosts x 6 rotating disks/host, usual mons, mgrs, mds): The
first time after a completely cold ceph cluster start (waiting for
health ok, idle otherwise): The command to create a ceph pool hangs,
I appreciate your efforts. The thing is folks who use freeipa put it in
the same 'has-got-to-work' 'no-regressions' category as the kernel.
While it might lack a feature or need work in this or that area, it just
can't 'not install' or have some major user-facing thing like the
'here's how you
Timo,
Thank you. I didn't understand freeipa wasn't supported on Ubuntu. You
can consider this matter closed, I have to move to a different distro.
On 11/25/19 2:20 PM, Timo Aaltonen wrote:
> replica install is untested, not surprising to see it being broken
>
> and freeipa is in universe
Of some interest, a curl of exactly the same link works (kinit admin in effect,
just after failure above).
root@registry2:/tmp# curl
Here's the shell script log
root@registry2:~# kinit admin
Password for ad...@1.quietfountain.com:
root@registry2:~# ipa-replica-install --setup-dns --no-forwarders
WARNING: conflicting time synchronization service 'ntp' will
be disabled in favor of chronyd
Lookup failed: Preferred host
Both registry1 and registry2 are 'vanilla' eoan mate vms.
Host registry1... has a working freeipa-server based on eoan installed. No
other packages. It does include the dns support. registry2 is the attempt to
install a replica. No other packages.
--
You received this bug notification
Public bug reported:
Just trying to see if freeipa works on Ubuntu, I installed freeipa-
server on one system, then tried to install the freeipa-replica on
another. The two system setup works just fine on Fedora, but I need to
standardize on one distro so I'm evaluating Ubuntu hoping that
Public bug reported:
Notice the bug and fix mentioned in
https://bugs.launchpad.net/ubuntu/+source/freeipa/+bug/1772921
is, somehow 're-broken' in eoan. Possibly because of:
https://pagure.io/freeipa/c/78652a52f083bac5238f9e0a6520e0e448dadabe
The result is none of the directional glyphs
Adding
usr/lib/*/ctdb/ctdb_mutex_ceph_rados_helper
and
usr/share/man/man7/ctdb_mutex_ceph_rados_helper.7
to
/debian/ctdb.install
puts those files in the ctdb deb and corrects the above.
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
When --enable-ceph-relock is added to rules, and librados-dev is installed,
dpkg-buildpackage reports this error:
ake[1]: Entering directory '/tmp/samba/samba-4.10.7+dfsg'
dh_missing --fail-missing
dh_missing: usr/lib/x86_64-linux-gnu/ctdb/ctdb_mutex_ceph_rados_helper exists
in debian/tmp but is
Public bug reported:
In the samba source subdirectory ctdb/utils/ceph there is
ctdb_mutex_ceph_rados_helper.c, needed for proper operation of ctdb in a
ceph environment. But, ubuntu doesn't include the --enable-ceph-reclock
option so that program doesn't appear in the ubuntu version of the
Public bug reported:
Would ubuntu include the nautilus patch
https://github.com/ceph/ceph/pull/28869 ?
Right now the ceph dashboard won't report 'normal'/'green' because of the
bug/hack of resetting scrub timestamps to 0 when auto-repair is set. The above
patch, accepted in May by Ceph, fixes
I see the same thing related to slowness in a old-style parallel port device
passed to a VM. Shortly after there's traffic related to the parallel port the
task will hang forever, eventually locking up the VM entirely.
This has happened on two SuperMicro servers, the problem did not exist in
deleting 10_linux_zfs restored previous normal zfs boot operations.
Note the script shows more severe errors than the extra }, the initrd arguments
were blank.
Perhaps some guidance for those of us that already have zfs roots as per
the previous guides... to have the new 'for desktops' feature
Public bug reported:
Despite the attempted patch to prevent a double } } by 10_linux_zfs in grub.cfg
for pure zfs systems, the failure remains in the latest eoan.
Total install failure results crashing do-release-upgrade. Critical.
A debug run follows, then the generated file with the extra }
Public bug reported:
I'm testing out eoan, I have a [Network] section with IPMasquerade=yes.
The system starts without complaint from systemd in the logs, but I see no
evidence of source natting going on.
I'm not sure of sub-dependencies, but I think source natting depends on
libiptc-dev, which
Upstream has two approaches to a solution. One was to disable sdpk
except for development versions because the spdk folks set their lowest
usable software level to corei7. I couldn't get that patch to work in
the ubuntu packaging 'apt-get source ceph'.
I was able to get the patch working that
Not so great minds think alike. Here it is, from upstream:
https://tracker.ceph.com/issues/41330
** Bug watch added: tracker.ceph.com/issues #41330
http://tracker.ceph.com/issues/41330
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to
I think I found it, but could use some validation.
Notice at
https://ceph.io/geen-categorie/sse-optimization-for-erasure-code-in-ceph/
we have the precedent for ceph's checking what level of SSE instructions are
available then using the appropriate one.
However, in the ubuntu version, littered
And with debug symbols:
(gdb) run
Starting program: /usr/bin/ceph-bluestore-tool
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
Program received signal SIGILL, Illegal instruction.
0x55743984 in
Try it setting the processors on the VM to dual conroe.
On 8/30/19 3:38 AM, Trent Lloyd wrote:
> At a super basic level I can't reproduce this. With an eoan container on
> an eoan host I don't get a segfault from ceph-bluestore-tool.
>
> I'd suggest we may need to look at getting
> (1) a
here's a core.gz of a typical crash in case the previous apport thing
didn't get it to you.
** Attachment added: "core.gz of typical crash"
https://bugs.launchpad.net/ubuntu/+source/ceph/+bug/1842020/+attachment/5285838/+files/core.gz
--
You received this bug notification because you are a
apport information
** Attachment added: "modified.conffile..etc.default.apport.txt"
https://bugs.launchpad.net/bugs/1842020/+attachment/5285824/+files/modified.conffile..etc.default.apport.txt
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed
apport information
** Tags added: apport-collected eoan
** Description changed:
The ceph patch on eoan distributed 8/29 crashes all bluestore OSD's and
is unusable at least on some systems. Failed on an old dual Xeon E5345
box.
Easy test. Run:
/usr/bin/ceph-bluestore-tool
apport information
** Attachment added: "ProcCpuinfoMinimal.txt"
https://bugs.launchpad.net/bugs/1842020/+attachment/5285823/+files/ProcCpuinfoMinimal.txt
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
FYI, in the eoan VM you used to duplicate the bug, try it again with the
processor set to 2 'Conroe' cpus. i440FX chipset, BIOS, kvm-spice
emulator.
Also, I tried compiling ceph nautilus from upstream, can't do it in eoan
without installing tox from bionic, and a couple other packages not in
eoan
The above data is from a run in an eoan VM on the same processor that hosts the
osd's in the baremetal layer (same bug in both cases, but I reverted the
baremetal layer to the state prior to the 'upgrade' to restore ceph osd
function).
Here's the dpkg -l you asked for.
Here's cat /proc/cpuinfo
Backtrace log for you.
** Attachment added: "backtrace log"
https://bugs.launchpad.net/ubuntu/+source/ceph/+bug/1842020/+attachment/5285808/+files/gdb-ceph-bluestore-tool.txt
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
Attached is the result of
apport-bug --save
it can be viewed with
apport-unpack
It has the answers to most of the above questions. Yes, it is amd64 (dual
xeon...)
** Attachment added: "result of apport-bug"
Bisected the problem starts with 14.2.2-0ubuntu1, might be the next one.
It works in 14.2.1-0ubuntu3. Look for the change in the file size of
ceph-bluestore-osd.
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
Confirmed on two different boxes of the same processor vintage.
Otherwise latest eoan updates. Desktop ceph mate though it shouldn't
matter.
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1842020
1 - 100 of 168 matches
Mail list logo
