Debian Bug report logs - #781007
network-manager has an 128 max connections limit

Toggle useless messages

Message #5 received at submit@bugs.debian.org (full text, mbox, reply):

From: Faidon Liambotis <paravoid@debian.org>

To: Debian Bug Tracking System <submit@bugs.debian.org>

Subject: network-manager has an 128 max connections limit

Date: Mon, 23 Mar 2015 07:37:26 +0200

Package: network-manager
Version: 0.9.10.0-6
Severity: important

OK, this may sound weird but...

I have noticed some odd behavior, with Network Manager not showing one
of my GSM providers on the tray icon at random. Then today, I wanted to
edit one of my connections to change the DNS server list and hit "Edit
connections" (which spawns nm-connection-editor) and could not find it.

Repeated invocations of nm-connection-editor showed different
connections being listed -- I even got the connection I wanted at some
point. I initially thought it was an NM GUI bug, so I tried both nmtui
and nmcli and the same thing happened: every time I called them I could
see a list of connections but somewhat different every time.

Then... I thought of counting them: it seems that "nmcli c |wc -l"
consistently reports back "129" (header line + 128 connections). Saving
+ diffing the output, though, seems to indicate that I get a different
list of 128 saved connections every time I call it (I have lots of them
as this includes a lot of random hotspots I've picked up while
travelling).

This is quite an annoying limitation and manifests quite badly in a hard
to debug way. I could see this being triaged as serious as well -- the
only reason I didn't is that probably not a lot of people have > 128
connections saved.

I'd prefer if I didn't list my connections here for privacy reasons, but
do let me know if there's anything you want to run on the system itself
for debugging purposes.

Thanks,
Faidon

Message #10 received at 781007@bugs.debian.org (full text, mbox, reply):

From: Michael Biebl <biebl@debian.org>

To: Faidon Liambotis <paravoid@debian.org>, 781007@bugs.debian.org

Subject: Re: [Pkg-utopia-maintainers] Bug#781007: network-manager has an 128 max connections limit

Date: Mon, 23 Mar 2015 11:30:50 +0100

[Message part 1 (text/plain, inline)]

Control: forcemerge 773525 -1

Am 23.03.2015 um 06:37 schrieb Faidon Liambotis:
> 
> I have noticed some odd behavior, with Network Manager not showing one
> of my GSM providers on the tray icon at random. Then today, I wanted to
> edit one of my connections to change the DNS server list and hit "Edit
> connections" (which spawns nm-connection-editor) and could not find it.
> 
> Repeated invocations of nm-connection-editor showed different
> connections being listed -- I even got the connection I wanted at some
> point. I initially thought it was an NM GUI bug, so I tried both nmtui
> and nmcli and the same thing happened: every time I called them I could
> see a list of connections but somewhat different every time.
> 
> Then... I thought of counting them: it seems that "nmcli c |wc -l"
> consistently reports back "129" (header line + 128 connections). Saving
> + diffing the output, though, seems to indicate that I get a different
> list of 128 saved connections every time I call it (I have lots of them
> as this includes a lot of random hotspots I've picked up while
> travelling).
> 
> This is quite an annoying limitation and manifests quite badly in a hard
> to debug way. I could see this being triaged as serious as well -- the
> only reason I didn't is that probably not a lot of people have > 128
> connections saved.
> 
> I'd prefer if I didn't list my connections here for privacy reasons, but
> do let me know if there's anything you want to run on the system itself
> for debugging purposes.

Looks like a duplicate of #773525.
Please test, if the latest upstream version 1.0.0 (available from
experimental) fixes the problem. If not, please file a bug upstream.

Thanks,
Michael



-- 
Why is it that all of the instruments seeking intelligent life in the
universe are pointed away from Earth?

[signature.asc (application/pgp-signature, attachment)]

Message #21 received at 781007@bugs.debian.org (full text, mbox, reply):

From: Faidon Liambotis <paravoid@debian.org>

To: Michael Biebl <biebl@debian.org>

Cc: 781007@bugs.debian.org

Subject: Re: [Pkg-utopia-maintainers] Bug#781007: network-manager has an 128 max connections limit

Date: Mon, 23 Mar 2015 14:49:53 +0200

On Mon, Mar 23, 2015 at 11:30:50AM +0100, Michael Biebl wrote:
> Looks like a duplicate of #773525.
> Please test, if the latest upstream version 1.0.0 (available from
> experimental) fixes the problem. If not, please file a bug upstream.

Doh -- sorry for the poor bug reporting.

I hadn't noticed that experimental had a newer version. I just checked
and it seems to be fixed there, output is now consistently the same list
of 182 connections.

I checked the git log between those two versions and didn't find
anything that immediately stands out. Bisecting will be 7 steps, so if
you have any guesses, I'm all ears... I'm guessing that if I do manage
to track it down, you'd consider it for jessie, right?

Thanks,
Faidon

Message #26 received at 781007@bugs.debian.org (full text, mbox, reply):

From: Michael Biebl <biebl@debian.org>

To: Faidon Liambotis <paravoid@debian.org>, 781007@bugs.debian.org

Subject: Re: [Pkg-utopia-maintainers] Bug#781007: Bug#781007: network-manager has an 128 max connections limit

Date: Mon, 23 Mar 2015 16:21:32 +0100

[Message part 1 (text/plain, inline)]

Am 23.03.2015 um 13:49 schrieb Faidon Liambotis:
> On Mon, Mar 23, 2015 at 11:30:50AM +0100, Michael Biebl wrote:
>> Looks like a duplicate of #773525.
>> Please test, if the latest upstream version 1.0.0 (available from
>> experimental) fixes the problem. If not, please file a bug upstream.
> 
> Doh -- sorry for the poor bug reporting.
> 
> I hadn't noticed that experimental had a newer version. I just checked
> and it seems to be fixed there, output is now consistently the same list
> of 182 connections.
> 
> I checked the git log between those two versions and didn't find
> anything that immediately stands out. Bisecting will be 7 steps, so if
> you have any guesses, I'm all ears... I'm guessing that if I do manage
> to track it down, you'd consider it for jessie, right?

Depends on the patch and how invasive it is.
I do indeed plan another upload of NetworkManager to at least fix
#760998. So we might sneak that one in.

Michael
-- 
Why is it that all of the instruments seeking intelligent life in the
universe are pointed away from Earth?

[signature.asc (application/pgp-signature, attachment)]

Message #31 received at 781007@bugs.debian.org (full text, mbox, reply):

From: Michael Biebl <biebl@debian.org>

To: Faidon Liambotis <paravoid@debian.org>, 781007@bugs.debian.org

Subject: Re: [Pkg-utopia-maintainers] Bug#781007: Bug#781007: network-manager has an 128 max connections limit

Date: Mon, 23 Mar 2015 16:22:32 +0100

[Message part 1 (text/plain, inline)]

Control: fixed -1 1.0.0-3

Am 23.03.2015 um 13:49 schrieb Faidon Liambotis:
> On Mon, Mar 23, 2015 at 11:30:50AM +0100, Michael Biebl wrote:
>> Looks like a duplicate of #773525.
>> Please test, if the latest upstream version 1.0.0 (available from
>> experimental) fixes the problem. If not, please file a bug upstream.
> 
> Doh -- sorry for the poor bug reporting.
> 
> I hadn't noticed that experimental had a newer version. I just checked
> and it seems to be fixed there, output is now consistently the same list
> of 182 connections.

Thanks for testing, marking the bug as fixed for that version


-- 
Why is it that all of the instruments seeking intelligent life in the
universe are pointed away from Earth?

[signature.asc (application/pgp-signature, attachment)]

Message #38 received at 781007@bugs.debian.org (full text, mbox, reply):

From: Faidon Liambotis <paravoid@debian.org>

To: Michael Biebl <biebl@debian.org>

Cc: 781007@bugs.debian.org

Subject: Re: [Pkg-utopia-maintainers] Bug#781007: Bug#781007: network-manager has an 128 max connections limit

Date: Tue, 24 Mar 2015 17:16:11 +0200

On Mon, Mar 23, 2015 at 04:21:32PM +0100, Michael Biebl wrote:
> Depends on the patch and how invasive it is.
> I do indeed plan another upload of NetworkManager to at least fix
> #760998. So we might sneak that one in.

OK, I managed to track this down further. After many failed attempts, I
had to script the bisect entirely (bisect run) during which I discovered
that the bug does not manifest if you call nmcli as root(!). IOW,

$ nmcli c | wc -l
129
$ sudo nmcli c |wc -l
183

The bisect ends up at 57e802f3aa2aac8daba9a7a442a2393ab5888dba ("libnm:
port NMRemoteSettings to NMObject") as the first "bad" commit (i.e. the
commit that fixed this). Unfortunately this is pretty invasive :(

Any clues on what specifically there would make such a difference? I'm
suspecting a DBus timeout somewhere...

For posterity, the full bisect log is:
# bad: [acdaf78a068b6c65ba799a7098b867953db4801c] release: bump version to 1.0.0
# good: [6eb82acd6dce882f4b91aafcf68dd9e143ce34e2] release: bump version to 0.9.10.0
git bisect start '1.0.0' '0.9.10.0'
# good: [c817409fc09044db9a871859c98e121074ea2495] release: bump version to 0.9.9.95 (0.9.10-beta1)
git bisect good c817409fc09044db9a871859c98e121074ea2495
# bad: [d83881a64125cd7d398834eec6a31ce7ee8d006f] tui, cli: improve printing of Wi-Fi strength bars (rh #1131491)
git bisect bad d83881a64125cd7d398834eec6a31ce7ee8d006f
# good: [75d8a805e76204d44235fcf750fee0a65313c419] core: add logging macro _LOG() and _LOGD() to nm-device.c
git bisect good 75d8a805e76204d44235fcf750fee0a65313c419
# bad: [290598494d46231a36a3982c943ad09eceb70eae] libnm-core: split the test program out of nm-param-spec-specialized.c
git bisect bad 290598494d46231a36a3982c943ad09eceb70eae
# bad: [8fe1b790126f433e8b9d05b3385e7075b6af68ae] libnm-core: declare NM_SETTING_COMPARE_FLAG_INFERRABLE flag in "nm-core-internal.h"
git bisect bad 8fe1b790126f433e8b9d05b3385e7075b6af68ae
# good: [7e791ce3fe250190db9a54ce95490e90bc6e26b3] libnm, libnm-glib: belatedly replace _nm_client_is_object_path()
git bisect good 7e791ce3fe250190db9a54ce95490e90bc6e26b3
# bad: [32c26a859bd869a5a910c5047685bcd2badce6d2] libnm-core: move some fake NMConnection methods over to NMSetting
git bisect bad 32c26a859bd869a5a910c5047685bcd2badce6d2
# bad: [df2711f1975446b42688c89db80e3fec105d1932] nmtst: fix memory leak in nmtst_create_minimal_connection()
git bisect bad df2711f1975446b42688c89db80e3fec105d1932
# bad: [f8762f7d3f0f916057f2c542163a02a77a6c3bbd] libnm: remove NMRemoteConnection::updated signal
git bisect bad f8762f7d3f0f916057f2c542163a02a77a6c3bbd
# good: [c4a86eba52e4665159546293a5065bab12077b69] libnm: remove redundant NM name watching code
git bisect good c4a86eba52e4665159546293a5065bab12077b69
# bad: [57e802f3aa2aac8daba9a7a442a2393ab5888dba] libnm: port NMRemoteSettings to NMObject
git bisect bad 57e802f3aa2aac8daba9a7a442a2393ab5888dba
# good: [d7e99f837537abc45b193d6748888f264e02d0b6] libnm: add NMRemoteConnection:visible property
git bisect good d7e99f837537abc45b193d6748888f264e02d0b6
# first bad commit: [57e802f3aa2aac8daba9a7a442a2393ab5888dba] libnm: port NMRemoteSettings to NMObject

Thanks,
Faidon

Message #43 received at 781007@bugs.debian.org (full text, mbox, reply):

From: Michael Biebl <biebl@debian.org>

To: Faidon Liambotis <paravoid@debian.org>, 781007@bugs.debian.org

Subject: Re: [Pkg-utopia-maintainers] Bug#781007: Bug#781007: Bug#781007: network-manager has an 128 max connections limit

Date: Tue, 24 Mar 2015 16:28:11 +0100

[Message part 1 (text/plain, inline)]

Am 24.03.2015 um 16:16 schrieb Faidon Liambotis:
> On Mon, Mar 23, 2015 at 04:21:32PM +0100, Michael Biebl wrote:
>> Depends on the patch and how invasive it is.
>> I do indeed plan another upload of NetworkManager to at least fix
>> #760998. So we might sneak that one in.
> 
> OK, I managed to track this down further. After many failed attempts, I
> had to script the bisect entirely (bisect run) during which I discovered
> that the bug does not manifest if you call nmcli as root(!). IOW,
> 

Just out of interest, if you check the stable branch, i.e. nm-0-9-10
from git, does that fix your issue as well.
0.9.10.2 does have quite a large diff, but it also has an impressive
long list of bug fixes.



-- 
Why is it that all of the instruments seeking intelligent life in the
universe are pointed away from Earth?

[signature.asc (application/pgp-signature, attachment)]

Message #48 received at 781007@bugs.debian.org (full text, mbox, reply):

From: Faidon Liambotis <paravoid@debian.org>

To: Michael Biebl <biebl@debian.org>

Cc: 781007@bugs.debian.org

Subject: Re: [Pkg-utopia-maintainers] Bug#781007: Bug#781007: Bug#781007: network-manager has an 128 max connections limit

Date: Tue, 24 Mar 2015 17:32:33 +0200

On Tue, Mar 24, 2015 at 04:28:11PM +0100, Michael Biebl wrote:
> Just out of interest, if you check the stable branch, i.e. nm-0-9-10
> from git, does that fix your issue as well.
> 0.9.10.2 does have quite a large diff, but it also has an impressive
> long list of bug fixes.

I just checked: it doesn't, bug is still there.

Faidon

Message #53 received at 781007@bugs.debian.org (full text, mbox, reply):

From: Michael Biebl <biebl@debian.org>

To: Faidon Liambotis <paravoid@debian.org>, 781007@bugs.debian.org

Subject: Re: [Pkg-utopia-maintainers] Bug#781007: Bug#781007: Bug#781007: network-manager has an 128 max connections limit

Date: Mon, 30 Mar 2015 23:42:24 +0200

[Message part 1 (text/plain, inline)]

Am 24.03.2015 um 16:16 schrieb Faidon Liambotis:
> On Mon, Mar 23, 2015 at 04:21:32PM +0100, Michael Biebl wrote:
>> Depends on the patch and how invasive it is.
>> I do indeed plan another upload of NetworkManager to at least fix
>> #760998. So we might sneak that one in.
> 
> OK, I managed to track this down further. After many failed attempts, I
> had to script the bisect entirely (bisect run) during which I discovered
> that the bug does not manifest if you call nmcli as root(!). IOW,
> 
> $ nmcli c | wc -l
> 129
> $ sudo nmcli c |wc -l
> 183
> 
> The bisect ends up at 57e802f3aa2aac8daba9a7a442a2393ab5888dba ("libnm:
> port NMRemoteSettings to NMObject") as the first "bad" commit (i.e. the
> commit that fixed this). Unfortunately this is pretty invasive :(
> 

I decided to not backport this fix, but I'll provide a 1.0.0 jesie
backport early on in the stretch development cycle.

Sorry for the inconvenience

Michael

-- 
Why is it that all of the instruments seeking intelligent life in the
universe are pointed away from Earth?

[signature.asc (application/pgp-signature, attachment)]

Message #54 received at 773525-done@bugs.debian.org (full text, mbox, reply):

From: Michael Biebl <biebl@debian.org>

To: 773525-done@bugs.debian.org

Subject: Re: Bug#773525: Randomly excludes available connections [when there are too many?]

Date: Fri, 3 Jun 2016 11:25:14 +0200

[Message part 1 (text/plain, inline)]

Version: 1.2.0-1

On Wed, 12 Aug 2015 11:58:36 -0500 Dan Williams <dcbw@redhat.com> wrote:
> On Fri, 2015-07-17 at 00:35 +0100, Simon McVittie wrote:
> > On 17/07/15 00:13, Dan Williams wrote:
> > > eg, think of it as each contact in Telepathy being an object
> 
> <snip>
> 
> > > The solution here is likely to transition the libnm implementation over
> > > to the ObjectManager interface's GetManagedObjects() method for the
> > > initial setup, to get everything in one call.  A lot of data, but likely
> > > faster than doing it piece-by-piece.
> > 
> > That's what I'd recommend; ObjectManager is precisely for situations
> > where clients are interested in "most" children of a parent object.
> 
> That's the path I'd like to pursue, I did an initial implementation of
> the OM interface for NM git master (what will become 1.2) based on our
> gdbus conversion (which is now merged! hurrah!!!) and filed a bug to
> track its ongoing work:
> 
> https://bugzilla.gnome.org/show_bug.cgi?id=753566


Version 1.2 is in unstable/testing now, Closing the bug accordingly.

Regards,
Michael
-- 
-- 
Why is it that all of the instruments seeking intelligent life in the
universe are pointed away from Earth?

[signature.asc (application/pgp-signature, attachment)]

Send a report that this bug log contains spam.