Re: user folder

[Gabriele Bulfon <gbulfon@xxxxxxxxxxx>]

Hi,
as suggested, I upgraded to 2.4.12, but master seems to hang on first telnet to 143, never
giving me the cyrus imap welcome string.

I tried to step back up to 2.4.10 always with the same problem.
Then got back to 2.4.2 and is fine again.

I used strace to check what master is doing on my first connect, and compared what happens
on 2.4.2 and on 2.4.12. Looks like 2.4.12 goes into a loop.
Here are the outputs of strace:

2.4.2
=====
[sonicle@xstream ~]$ sudo strace -p 19790
Process 19790 attached - interrupt to quit
select(33, [7 8 10 11 13 14 16 17 19 20 22 23 25 26 28 29 31 32], NULL, NULL, {1, 95000}) = 0 (Timeout)
time(NULL) = 1318492904
gettimeofday({1318492904, 52043}, NULL) = 0
select(33, [7 8 10 11 13 14 16 17 19 20 22 23 25 26 28 29 31 32], NULL, NULL, {10, 0}) = 1 (in [10], left {1, 804000})
time(NULL) = 1318492912
clone(child_stack=0, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|SIGCHLD, child_tidptr=0xb7f9d708) = 19802
time(NULL) = 1318492912
gettimeofday({1318492912, 251845}, NULL) = 0
select(33, [7 8 11 13 14 16 17 19 20 22 23 25 26 28 29 31 32], NULL, NULL, {2, 0}) = 1 (in [11], left {1, 938000})
read(11, "\2\0\0\0ZM\0\0", 8) = 8
read(11, 0xbf833a34, 8) = -1 EAGAIN (Resource temporarily unavailable)
time(NULL) = 1318492912
gettimeofday({1318492912, 314802}, NULL) = 0
select(33, [7 8 10 11 13 14 16 17 19 20 22 23 25 26 28 29 31 32], NULL, NULL, {2, 0}) = 1 (in [11], left {2, 0})
read(11, "\3\0\0\0ZM\0\0", 8) = 8
read(11, 0xbf833a34, 8) = -1 EAGAIN (Resource temporarily unavailable)
time(NULL) = 1318492912
gettimeofday({1318492912, 315531}, NULL) = 0
select(33, [7 8 10 11 13 14 16 17 19 20 22 23 25 26 28 29 31 32], NULL, NULL, {2, 0}) = 0 (Timeout)
time(NULL) = 1318492914
gettimeofday({1318492914, 315843}, NULL) = 0
select(33, [7 8 10 11 13 14 16 17 19 20 22 23 25 26 28 29 31 32], NULL, NULL, {10, 0} <unfinished ...>
......

2.4.12
======
select(33, [7 8 10 11 13 14 16 17 19 20 22 23 25 26 28 29 31 32], NULL, NULL, {8, 0}) = 1 (in [10], left {8, 0})
time(NULL) = 1318493061
clone(child_stack=0, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|SIGCHLD, child_tidptr=0xb7fef708) = 20196
time(NULL) = 1318493061
gettimeofday({1318493061, 990811}, NULL) = 0
select(33, [7 8 11 13 14 16 17 19 20 22 23 25 26 28 29 31 32], NULL, NULL, {8, 0}) = ? ERESTARTNOHAND (To be restarted)
--- SIGCHLD (Child exited) @ 0 (0) ---
sigreturn() = ? (mask now [])
waitpid(-1, [{WIFEXITED(s) && WEXITSTATUS(s) == 75}], WNOHANG) = 20196
time(NULL) = 1318493061
stat64("/etc/localtime", {st_mode=S_IFREG|0644, st_size=2652, ...}) = 0
stat64("/etc/localtime", {st_mode=S_IFREG|0644, st_size=2652, ...}) = 0
stat64("/etc/localtime", {st_mode=S_IFREG|0644, st_size=2652, ...}) = 0
send(5, "<183>Oct 13 10:04:21 master[2015"..., 67, MSG_NOSIGNAL) = 67
time(NULL) = 1318493061
stat64("/etc/localtime", {st_mode=S_IFREG|0644, st_size=2652, ...}) = 0
stat64("/etc/localtime", {st_mode=S_IFREG|0644, st_size=2652, ...}) = 0
stat64("/etc/localtime", {st_mode=S_IFREG|0644, st_size=2652, ...}) = 0
send(5, "<180>Oct 13 10:04:21 master[2015"..., 96, MSG_NOSIGNAL) = 96
time(NULL) = 1318493061
waitpid(-1, 0xbfb42b64, WNOHANG) = -1 ECHILD (No child processes)
select(33, [7 8 10 11 13 14 16 17 19 20 22 23 25 26 28 29 31 32], NULL, NULL, {8, 0}) = 1 (in [10], left {8, 0})
time(NULL) = 1318493061
clone(child_stack=0, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|SIGCHLD, child_tidptr=0xb7fef708) = 20197
time(NULL) = 1318493061
gettimeofday({1318493062, 264}, NULL) = 0
select(33, [7 8 11 13 14 16 17 19 20 22 23 25 26 28 29 31 32], NULL, NULL, {8, 0}) = ? ERESTARTNOHAND (To be restarted)
--- SIGCHLD (Child exited) @ 0 (0) ---
sigreturn() = ? (mask now [])
waitpid(-1, [{WIFEXITED(s) && WEXITSTATUS(s) == 75}], WNOHANG) = 20197
time(NULL) = 1318493062
stat64("/etc/localtime", {st_mode=S_IFREG|0644, st_size=2652, ...}) = 0
stat64("/etc/localtime", {st_mode=S_IFREG|0644, st_size=2652, ...}) = 0
stat64("/etc/localtime", {st_mode=S_IFREG|0644, st_size=2652, ...}) = 0
send(5, "<183>Oct 13 10:04:22 master[2015"..., 67, MSG_NOSIGNAL) = 67
time(NULL) = 1318493062
stat64("/etc/localtime", {st_mode=S_IFREG|0644, st_size=2652, ...}) = 0
stat64("/etc/localtime", {st_mode=S_IFREG|0644, st_size=2652, ...}) = 0
stat64("/etc/localtime", {st_mode=S_IFREG|0644, st_size=2652, ...}) = 0
send(5, "<180>Oct 13 10:04:22 master[2015"..., 96, MSG_NOSIGNAL) = 96
time(NULL) = 1318493062
waitpid(-1, 0xbfb42b64, WNOHANG) = -1 ECHILD (No child processes)
select(33, [7 8 10 11 13 14 16 17 19 20 22 23 25 26 28 29 31 32], NULL, NULL, {8, 0}) = 1 (in [10], left {8, 0})
time(NULL) = 1318493062
clone(child_stack=0, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|SIGCHLD, child_tidptr=0xb7fef708) = 20198
time(NULL) = 1318493062
gettimeofday({1318493062, 9412}, NULL) = 0

never ending loop here.

Any clue?

----------------------------------------------------------------------------------

Da: Bron Gondwana <brong@xxxxxxxxxxx>
A: Gabriele Bulfon <gbulfon@xxxxxxxxxxx>
Cc: info-cyrus@xxxxxxxxxxxxxxxxxxxx
Data: 11 ottobre 2011 15.48.54 CEST
Oggetto: Re: user folder

Date: Tue Oct 19 14:08:28 2010 +1100

Release 2.4.2

There were numerous bugs in early 2.4 versions related to configurations
which I hadn't used, and which hadn't been tested enough when developing
it. As others have said, 2.4.2 is very early 2.4 series, and had many
bugs. 2.4.12 is much better.

Bron.


On 10/11/2011 11:57 AM, Gabriele Bulfon wrote:
> Hi
>
> I'm using cyrus in production for years, recently I'm working on a new
> distribution with a new
> update cyrus, built from sources.
> Previous version was 2.3.16, new version is 2.4.2.
>
> On the previous version, you could issue 'LIST "" user' to the imap
> server and receive this response:
>
> A0 LIST "" user
> * LIST (\Noselect \HasChildren) "." "user"
> A0 OK Completed (0.260 secs 419 calls)
>
> In the new version the response is different:
>
> A0 LIST "" user
> A0 OK Completed (0.030 secs 202 calls)
>
> This open up a problem to our web app using JavaMail, when it tries to
> list the shared folders from
> other users.
> The code examines the user namespaces and finds "user".
> Then it asks JavaMail to list the folder "user" to obtain the list of
> users sharing something.
> Debugging the IMAP transaction, I see JavaMail trying to do the above
> command first,
> then, after receiving the response, it will issue a LIST "" user.% and
> go on.
> Now, because the above command answers with nothing, I can't seem to
> find a correct way to
> list the "user" magic folder.
>
> Why has this changed in Cyrus?
> Is there any way I can let it work as before?
>
> Thanks,
> Gabriele.
>
>
> ----
> Cyrus Home Page: http://www.cyrusimap.org/
> List Archives/Info: http://lists.andrew.cmu.edu/pipermail/info-cyrus/

----
Cyrus Home Page: http://www.cyrusimap.org/
List Archives/Info: http://lists.andrew.cmu.edu/pipermail/info-cyrus/