Filed follow-up
https://bugs.launchpad.net/ubuntu/+source/mariadb/+bug/2038500 to track
when the underlying issue is fixed and the workaround can be removed.
--
You received this bug notification because you are a member of Ubuntu
Touch seeded packages, which is subscribed to systemd in Ubuntu.
The commit mentioned in
https://bugs.launchpad.net/ubuntu/+source/systemd/+bug/1970634/comments/3
has been present in systemd since v251 (which is in Kinetic), so I don't
think there is anything left to do for systemd. Please correct me if I
am wrong.
** Changed in: systemd (Ubuntu)
Removing the systemd jammy task, it's unlikely such a change would be
SRUed.
** Changed in: systemd (Ubuntu Jammy)
Status: Confirmed => Won't Fix
** Tags removed: server-todo
--
You received this bug notification because you are a member of Ubuntu
Touch seeded packages, which is
This bug was fixed in the package mariadb-10.6 - 1:10.6.7-2ubuntu1.1
---
mariadb-10.6 (1:10.6.7-2ubuntu1.1) jammy; urgency=medium
* d/rules: disable LTO on Ubuntu so a low MEMLOCK_LIMIT can be
properly caught and dealt with (LP: #1970634)
-- Andreas Hasenack Fri, 17 Jun
Status changed to 'Confirmed' because the bug affects multiple users.
** Changed in: systemd (Ubuntu Jammy)
Status: New => Confirmed
--
You received this bug notification because you are a member of Ubuntu
Touch seeded packages, which is subscribed to systemd in Ubuntu.
Will also fix #1979695
--
You received this bug notification because you are a member of Ubuntu
Touch seeded packages, which is subscribed to systemd in Ubuntu.
https://bugs.launchpad.net/bugs/1970634
Title:
FTBFS: mariadb fails to start due to low MEMLOCK limit
Status in mariadb-10.6
** Changed in: systemd (Ubuntu Jammy)
Status: In Progress => New
** Changed in: systemd (Ubuntu Jammy)
Assignee: Andreas Hasenack (ahasenack) => (unassigned)
** Changed in: mariadb-10.6 (Ubuntu Jammy)
Status: Triaged => In Progress
** Description changed:
[Impact]
** Merge proposal linked:
https://code.launchpad.net/~ahasenack/ubuntu/+source/mariadb-10.6/+git/mariadb-10.6/+merge/424986
--
You received this bug notification because you are a member of Ubuntu
Touch seeded packages, which is subscribed to systemd in Ubuntu.
** Description changed:
[Impact]
Jammy's MariaDB will fail to build, and also fail to start, if the
underlying kernel is 5.4.x (focal's) and if it's running in an
unprivileged container (lxd, docker). It will also fail to build in
launchpad builders.
Common scenarios where this
** Description changed:
[Impact]
- * An explanation of the effects of the bug on users and
+ Jammy's MariaDB was built with io_uring support, and it tries to enable
+ it at runtime if it deems it's running on a supported kernel. There is a
+ range of kernel versions it checks, but of
** Description changed:
+ [Impact]
+
+ * An explanation of the effects of the bug on users and
+
+ * justification for backporting the fix to the stable release.
+
+ * In addition, it is helpful, but not required, to include an
+explanation of how the upload fixes this bug.
+
+ [Test
Ok, phew, my test was incorrect, the bug is present in jammy. That makes
this SRU simpler. I'll proceed with it, without a block-proposed tag.
--
You received this bug notification because you are a member of Ubuntu
Touch seeded packages, which is subscribed to systemd in Ubuntu.
I got a report on IRC that the existing jammy package of mariadb does
exhibit this problem:
(link will expire soon)
Ok, a plain jammy rebuild in a launchpad ppa failed:
2022-06-17 14:18:27 0 [Warning] mariadbd: io_uring_queue_init() failed
with ENOMEM: try larger memory locked limit, ulimit -l, or
https://mariadb.com/kb/en/systemd/#configuring-limitmemlock under
systemd (262144 bytes required)
The binary
I can't reproduce this bug anymore with the package currently in jammy
(1:10.6.7-2ubuntu1):
lxc launch ubuntu:focal f --vm
lxc shell f
lxd init # just hit enter for all questions
lxc launch ubuntu:jammy j
lxc shell j
ulimit -l # confirm it's less than 256
apt update && apt install mariadb-server
Hi Michal, I'm going over the test procedure for this bug and turns out
the scenario where we encountered it is a bit convoluted, specially for
users: you need to be running a 5.4 kernel on a jammy userspace (jammy
itself has 5.15 kernel).
Could you please elaborate on how this bug affects you?
** Changed in: systemd (Ubuntu Jammy)
Assignee: (unassigned) => Andreas Hasenack (ahasenack)
** Changed in: systemd (Ubuntu Jammy)
Status: New => In Progress
** Changed in: systemd (Ubuntu Jammy)
Importance: Undecided => High
--
You received this bug notification because you are
** Also affects: systemd (Ubuntu Jammy)
Importance: Undecided
Status: New
** Also affects: mariadb-10.6 (Ubuntu Jammy)
Importance: Undecided
Status: New
** Changed in: mariadb-10.6 (Ubuntu Jammy)
Assignee: (unassigned) => Andreas Hasenack (ahasenack)
** Changed in:
Hello, any idea when the fix will land in jammy?
--
You received this bug notification because you are a member of Ubuntu
Touch seeded packages, which is subscribed to systemd in Ubuntu.
https://bugs.launchpad.net/bugs/1970634
Title:
FTBFS: mariadb fails to start due to low MEMLOCK limit
This bug was fixed in the package mariadb-10.6 - 1:10.6.7-3ubuntu1
---
mariadb-10.6 (1:10.6.7-3ubuntu1) kinetic; urgency=medium
* d/rules: disable LTO (LP: #1970634)
-- Andreas Hasenack Fri, 29 Apr 2022 09:30:44
-0300
** Changed in: mariadb-10.6 (Ubuntu)
Status: In
** Merge proposal linked:
https://code.launchpad.net/~ahasenack/ubuntu/+source/mariadb-10.6/+git/mariadb-10.6/+merge/420845
--
You received this bug notification because you are a member of Ubuntu
Touch seeded packages, which is subscribed to systemd in Ubuntu.
> Could it be -flto/-ffat-lto-objects related (like
> https://jira.mariadb.org/browse/MDEV-25633)?
> The top part of the stack trace looks the same.
Nice catch. Indeed, disabling lto fixes the build and startup in low
memlock conditions.
I'm still concerned with lto creeping in via
Status changed to 'Confirmed' because the bug affects multiple users.
** Changed in: systemd (Ubuntu)
Status: New => Confirmed
--
You received this bug notification because you are a member of Ubuntu
Touch seeded packages, which is subscribed to systemd in Ubuntu.
Could it be -flto/-ffat-lto-objects related (like
https://jira.mariadb.org/browse/MDEV-25633)? The top part of the stack
trace looks the same.
--
You received this bug notification because you are a member of Ubuntu
Touch seeded packages, which is subscribed to systemd in Ubuntu.
On the assembly generated by create_linux_aio (in the mariadbd packages
in Ubuntu-22.04, 10.6.7-2:
(gdb) disassemble tpool::create_linux_aio,+200
Dump of assembler code from 0x5640cf2e1fb0 to 0x5640cf2e2078:
0x5640cf2e1fb0 <_ZN5tpool16create_linux_aioEPNS_11thread_poolEi+0>:
endbr64
Note the runtime detection in innodb_use_native_aio_default affects the
default value only. Users can explicity set this and really on distros
to provide patched kernels.
--
You received this bug notification because you are a member of Ubuntu
Touch seeded packages, which is subscribed to
Actually create_linux_aio (tpool/aio_liburing.cc:194) should already
catch this exception. So compiler error?
--
You received this bug notification because you are a member of Ubuntu
Touch seeded packages, which is subscribed to systemd in Ubuntu.
https://bugs.launchpad.net/bugs/1970634
Title:
Assertion cause:
$ podman run --rm --cap-add=CAP_SYS_PTRACE -e MARIADB_ROOT_PASSWORD=bob -ti
--user mysql m106_jammy_debug bash
mysql@0740c1895ab4:/$ gdb --args mariadbd --bootstrap
GNU gdb (Ubuntu 12.0.90-0ubuntu1) 12.0.90
Copyright (C) 2022 Free Software Foundation, Inc.
License GPLv3+:
>From 5.12 kernel memory locked pages aren't needed -
https://github.com/axboe/liburing/issues/246#issuecomment-816965961.
In #1969160 I tested the upcoming 10.6 release (probably next week
sometime) for crashes in uring initialization failures and it didn't.
Apart from some error log differences
Since mariadb on the current jammy kernel disables io_uring at startup,
I'm considering disabling io_uring entirely in the jammy mariadb build.
The only scenario where io_uring would be used by the jammy mariadb is
if the user ran a different kernel than the one shipped with jammy.
--
You
30 matches
Mail list logo
