Bug 799718
Summary: | [abrt] evolution-data-server-3.3.90-2.fc17: allocator_memalign: Process /usr/libexec/evolution-addressbook-factory was killed by signal 11 (SIGSEGV) | ||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Product: | [Fedora] Fedora | Reporter: | Nicolas Mailhot <nicolas.mailhot> | ||||||||||||
Component: | evolution-data-server | Assignee: | Matthew Barnes <mbarnes> | ||||||||||||
Status: | CLOSED DUPLICATE | QA Contact: | Fedora Extras Quality Assurance <extras-qa> | ||||||||||||
Severity: | unspecified | Docs Contact: | |||||||||||||
Priority: | unspecified | ||||||||||||||
Version: | rawhide | CC: | mbarnes, mcrha | ||||||||||||
Target Milestone: | --- | ||||||||||||||
Target Release: | --- | ||||||||||||||
Hardware: | x86_64 | ||||||||||||||
OS: | Unspecified | ||||||||||||||
Whiteboard: | abrt_hash:c765fed981dc7e0f90c2d8e31d919d50cd87f066 | ||||||||||||||
Fixed In Version: | Doc Type: | Bug Fix | |||||||||||||
Doc Text: | Story Points: | --- | |||||||||||||
Clone Of: | Environment: | ||||||||||||||
Last Closed: | 2012-03-30 13:08:35 UTC | Type: | --- | ||||||||||||
Regression: | --- | Mount Type: | --- | ||||||||||||
Documentation: | --- | CRM: | |||||||||||||
Verified Versions: | Category: | --- | |||||||||||||
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |||||||||||||
Cloudforms Team: | --- | Target Upstream Version: | |||||||||||||
Embargoed: | |||||||||||||||
Attachments: |
|
Description
Nicolas Mailhot
2012-03-04 16:15:01 UTC
Created attachment 567403 [details]
File: dso_list
Created attachment 567404 [details]
File: maps
Created attachment 567405 [details]
File: backtrace
Thanks for a bug report. I do not see any specific actions being processed inside the evolution-addressbook-factory process, all the calls are coming from GLib. The crashing thread, main thread, is currently processing socket_source_dispatch from gsocket.c, failing on memory allocation, which suggests there is some memory corruption. Was this just a "random" crash, or you can reproduce this reliably, please? It happens pretty reliably after reboot and log-in, and the kernel gets quickly confused Good, then maybe we can try to involve valgrind, which may help us to identify the memory corruption, if at all. Could you rename /usr/libexec/evolution-addressbook-factory to /usr/libexec/evolution-addressbook-factory.real and then create a script named /usr/libexec/evolution-addressbook-factory with the following content: #!/bin/bash G_SLICE=always-malloc valgrind --num-callers=50 /usr/libexec/evolution-addressbook-factory.real &>/tmp/eaflog.txt (note the line beginning with G_SLICE is one long line, it is not wrapped like here in bugzilla) then set executable attributes on the created script, like $ chmod a+x /usr/libexec/evolution-addressbook-factory and then try to reproduce the issue, please? If it'll crash again (though valgrind can avoid certain types of crashes and log about them only), then the file /tmp/eaflog.txt may contain what we are looking for. The file is overwritten each start of the factory, which can be started on demand by DBus, but we'll not clash here, hopefully. It crashed again but the logfile only contains EDataFactory is now online. Registering EBookBackendVCFFactory ('vcf') Registering EBookBackendLDAPFactory ('ldap') Registering EBookBackendGoogleFactory ('google') Registering EBookBackendWebdavFactory ('webdav') Registering EBookBackendFileFactory ('local') Server is up and running... Bus name 'org.gnome.evolution.dataserver.AddressBook3' acquired. libebookbackend-WARNING **: conversion to ldap query string failed libebookbackend-WARNING **: conversion to ldap query string failed THanks for giving it a try. It seems to me that the valgrind didn't make it, as the log should contain lines reported by it, same as the process itself is significantly slower, eating your CPU due to all memory checking. With > (note the line beginning with G_SLICE is one long line, it is not wrapped > like here in bugzilla) I meant that the line beginning with G_SLICE is not wrapped into two lines like in comment #6, but it should be one long line with below beginning and end: G_SLICE=.......factory.real &>/tmp/eaflog.txt aka make those two consecutive lines a one long line. I suppose this is the reason why valgrind is not involved here. Created attachment 568071 [details]
eaflog.txt
Sorry for not reading you correctly yesterday
Thanks for the update. The factory ends strange, it's saying: > Bus name 'org.gnome.evolution.dataserver.AddressBook3' lost. > g_dbus_connection_real_closed: Remote peer vanished with error: Underlying > GIOStream returned 0 bytes on an async read (g-io-error-quark, 0). Exiting. > Bye. which might mean other factory had been run over this one. The valgrind log also shows invalid reads I didn't notice yet, but they might be related to this error: > libebookbackend-WARNING **: conversion to ldap query string failed which suggests there is something wrong wit the expression being passed over the bus to the factory. The invalid reads are basically about the same, from the same part. Is there any special search activated in your Contacts view? You may see in in the top of the window, when you run evolution as "evolution -c contacts", unless you have evolution run in mailer and the addressbook is used to check whether the mail sender is in any of your books. (In reply to comment #10) > unless you have evolution run in mailer and the addressbook is used > to check whether the mail sender is in any of your books. I only use evolution as a mailer (and not that much anymore due to the bugs it accumulated in the past year) Created attachment 568384 [details]
contacts screenshot
Hmm, I'm still confused with the GIOStream error. Some of similar bug reports from [1] suggest this is related to GLib2 itself. I do not use the Fedora 17 as my main system yet, thus I do not have much more idea what to try. Just replace back the /usr/libexec/evolution-addressbook-factory to the real executable, to not have left there files after our testing. Is the Ekiga addressbook publicly available? I would like to try to connect to it, whether I'll be able to reproduce. But maybe this is reproducible with any LDAP addressbook. [1] https://bugzilla.gnome.org/buglist.cgi?query_format=advanced;longdesc=g_dbus_connection_real_closed;longdesc_type=allwordssubstr I tested this with my LDAP server and it works with no issue. I also see the warning:
> libebookbackend-WARNING **: conversion to ldap query string failed
but it has no influence on functionality. I see it when I clear previous search on the book, which is in a moment where you even cannot get due to the error of auto-closed addressbook factory process. I tried with up-to-date Fedora 17 installation.
(In reply to comment #13) > Is the Ekiga addressbook publicly available? I would like to try to connect to > it, whether I'll be able to reproduce. The ekiga adressbook is public and was set up by ekiga when I installed it https://www.ekiga.net/index.php?page=services https://admin.fedoraproject.org/pkgdb/acls/name/ekiga I suppose they add the addressbook to evolution only if I run ekiga with a registered user account, because I installed it and I do not have the Ekiga's LDAP addressbook added, though itself Addressbook (Ctrl+B) in Ekiga contains Ekiga's public LDAP book, the one, which doesn't require authentication. (In reply to comment #16) > I suppose they add the addressbook to evolution only if I run ekiga with a > registered user account, because I installed it and I do not have the Ekiga's > LDAP addressbook added, though itself Addressbook (Ctrl+B) in Ekiga contains > Ekiga's public LDAP book, the one, which doesn't require authentication. I don't know :( If you want to test, ekiga accounts are free Backtrace analysis found this bug to be similar to bug #759143, closing as duplicate. Bugs which were found to be similar to this bug: DeviceKit-disks: bug #579968 claws-mail: bug #696680, bug #749177, bug #789636 evince: bug #615380 evolution: bug #744469 evolution-data-server: bug #607960 evolution-exchange: bug #669649 gnome-shell: bug #759143, bug #781780 gvfs: bug #571349 pcmanfm: bug #789628 rhythmbox: bug #681163, bug #709834, bug #748661 xfce4-sensors-plugin: bug #772409 This comment is automatically generated. *** This bug has been marked as a duplicate of bug 759143 *** |