Mailing List Archive

Re: [DAViCal-devel] Possible fix for deadlocks in database
Marten,

I am encountering similar database deadlock issue as Egoitz described and
try to fix (here https://sourceforge.net/p/davical/mailman/message/34986381/
and here https://sourceforge.net/p/davical/mailman/message/35006191/) with
Postgres on Linux.

For testing with client, I used various version of Contacts.app on Mac OS X
10.9, 10.10, 10.11...

With the following config, no issue...
When I try to sync 5 contacts, five are effectively stored to the server...
the others are not...

Srl config:

AWL version 0.55
Davical version 1.1.3.1
psql (PostgreSQL) 9.3.12
PHP 5.5.9-1ubuntu4.16 (cli) (built: Apr 20 2016 14:31:27)
Copyright (c) 1997-2014 The PHP Group
Zend Engine v2.5.0, Copyright (c) 1998-2014 Zend Technologies
with Zend OPcache v7.0.3, Copyright (c) 1999-2014, by Zend Technologies
Distributor ID: Ubuntu
Description: Ubuntu 14.04.4 LTS
Release: 14.04
Codename: trusty
Server version: Apache/2.4.7 (Ubuntu)
Server built: Jan 14 2016 17:45:23

With the following config... the errors like the one further below appear...
When I try to sync 5 contacts, only one is stored to server... the others
are not...

Wood config:

AWL version 0.56
Davical version 1.1.4
psql (PostgreSQL) 9.1.21
contains support for command-line editing
PHP 5.3.10-1ubuntu3.22 with Suhosin-Patch (cli) (built: Apr 20 2016
12:31:15)
Copyright (c) 1997-2012 The PHP Group
Zend Engine v2.3.0, Copyright (c) 1998-2012 Zend Technologies
Distributor ID: Ubuntu
Description: Ubuntu 12.04.5 LTS
Release: 12.04
Codename: precise
Server version: Apache/2.2.22 (Ubuntu)
Server built: Jul 24 2015 17:25:42

Postgres error:
"2016-05-04 07:42:19 BST ERROR: deadlock detected
2016-05-04 07:42:19 BST DETAIL: Process 6162 waits for ShareLock on
transaction 192145; blocked by process 6164.
Process 6164 waits for ExclusiveLock on tuple (1,27) of relation
17248 of database 17112; blocked by process 6162.
Process 6162: INSERT INTO caldav_data ( user_no, dav_name,

dav_etag,

caldav_data, caldav_type, logged_user, created, modified,

collection_id )

VALUES( 1006,
'/richard/addresses/ac2fee95-ef58-451d-ae1d-58c3f6eca1b6.vcf',
'61413f34312ce41687ee8e4fc1a22c1f', E'BEGIN\x3aVCARD
(...)
"

That sounds likes there are 2 Apache processes (6162 and 6164) both trying
to access the same table ("caldav_data"). Process 6164 is waiting for
process 6162 to finish.

Issue raise with both PUT and DELETE commands...

No error reported on client during PUT sync...
HTTP 500 error reported on client which stop to sync during DELETE...

As in the first config case the issue does not raise, we believe it might
be fixable in Davical...

A short term fix we are trying is limiting the Apache connections for
Davical to only one per IP address (simplest but
not best solution), a Davical implemented solution may be best...

Any advice how to sort this issue?
Re: [DAViCal-devel] Possible fix for deadlocks in database [ In reply to ]
Hi,

Just to give you an update...

Upgrading Davical on our Srl config environment...

from:
AWL version 0.55
Davical version 1.1.3.1

to:
AWL version 0.56
Davical version 1.1.4

... confirmed that the issue is not linked to Davical but to Postgres or
else...

Indeed in both configuration in our Srl config environment we obtain the
same result, bulk PUT and DELETE complete successfully without Postgres
errors (deadlock) as previously encountered in our Woody config environment.

So a Distro upgrade (including a Postgres upgrade) seems to have sorted the
issue. Though in our case it is two separate environment (configured the
same way automatically) we did not upgraded as per say...

Now I have a network issue on Contacts.app when syncing ~5000 contacts!
It is never ending... getting there... getting there... :-)

Just though to keep you posted in case it may be useful to others.

Cheers,

On Wed, May 4, 2016 at 11:04 AM, Richard LEGER <richard.leger@gmail.com>
wrote:

>
> Marten,
>
> I am encountering similar database deadlock issue as Egoitz described and
> try to fix (here
> https://sourceforge.net/p/davical/mailman/message/34986381/ and here
> https://sourceforge.net/p/davical/mailman/message/35006191/) with
> Postgres on Linux.
>
> For testing with client, I used various version of Contacts.app on Mac OS
> X 10.9, 10.10, 10.11...
>
> With the following config, no issue...
> When I try to sync 5 contacts, five are effectively stored to the
> server... the others are not...
>
> Srl config:
>
> AWL version 0.55
> Davical version 1.1.3.1
> psql (PostgreSQL) 9.3.12
> PHP 5.5.9-1ubuntu4.16 (cli) (built: Apr 20 2016 14:31:27)
> Copyright (c) 1997-2014 The PHP Group
> Zend Engine v2.5.0, Copyright (c) 1998-2014 Zend Technologies
> with Zend OPcache v7.0.3, Copyright (c) 1999-2014, by Zend Technologies
> Distributor ID: Ubuntu
> Description: Ubuntu 14.04.4 LTS
> Release: 14.04
> Codename: trusty
> Server version: Apache/2.4.7 (Ubuntu)
> Server built: Jan 14 2016 17:45:23
>
> With the following config... the errors like the one further below
> appear...
> When I try to sync 5 contacts, only one is stored to server... the others
> are not...
>
> Wood config:
>
> AWL version 0.56
> Davical version 1.1.4
> psql (PostgreSQL) 9.1.21
> contains support for command-line editing
> PHP 5.3.10-1ubuntu3.22 with Suhosin-Patch (cli) (built: Apr 20 2016
> 12:31:15)
> Copyright (c) 1997-2012 The PHP Group
> Zend Engine v2.3.0, Copyright (c) 1998-2012 Zend Technologies
> Distributor ID: Ubuntu
> Description: Ubuntu 12.04.5 LTS
> Release: 12.04
> Codename: precise
> Server version: Apache/2.2.22 (Ubuntu)
> Server built: Jul 24 2015 17:25:42
>
> Postgres error:
> "2016-05-04 07:42:19 BST ERROR: deadlock detected
> 2016-05-04 07:42:19 BST DETAIL: Process 6162 waits for ShareLock on
> transaction 192145; blocked by process 6164.
> Process 6164 waits for ExclusiveLock on tuple (1,27) of relation
> 17248 of database 17112; blocked by process 6162.
> Process 6162: INSERT INTO caldav_data ( user_no, dav_name,
>
> dav_etag,
>
> caldav_data, caldav_type, logged_user, created, modified,
>
> collection_id )
>
> VALUES( 1006,
> '/richard/addresses/ac2fee95-ef58-451d-ae1d-58c3f6eca1b6.vcf',
> '61413f34312ce41687ee8e4fc1a22c1f', E'BEGIN\x3aVCARD
> (...)
> "
>
> That sounds likes there are 2 Apache processes (6162 and 6164) both trying
> to access the same table ("caldav_data"). Process 6164 is waiting for
> process 6162 to finish.
>
> Issue raise with both PUT and DELETE commands...
>
> No error reported on client during PUT sync...
> HTTP 500 error reported on client which stop to sync during DELETE...
>
> As in the first config case the issue does not raise, we believe it might
> be fixable in Davical...
>
> A short term fix we are trying is limiting the Apache connections for
> Davical to only one per IP address (simplest but
> not best solution), a Davical implemented solution may be best...
>
> Any advice how to sort this issue?
>
Re: [DAViCal-devel] Possible fix for deadlocks in database [ In reply to ]
Hi!!

For the deadlock issue you could use my patch. I use it in production without issues and fixes that problems.

Best regards,

Egoirz Aurrekoetxea Aurre
egoitz@ramattack.net
Sent from my smartphone

> El 4 may 2016, a las 19:45, Richard LEGER <richard.leger@gmail.com> escribió:
>
> Hi,
>
> Just to give you an update...
>
> Upgrading Davical on our Srl config environment...
>
> from:
> AWL version 0.55
> Davical version 1.1.3.1
>
> to:
> AWL version 0.56
> Davical version 1.1.4
>
> ... confirmed that the issue is not linked to Davical but to Postgres or else...
>
> Indeed in both configuration in our Srl config environment we obtain the same result, bulk PUT and DELETE complete successfully without Postgres errors (deadlock) as previously encountered in our Woody config environment.
>
> So a Distro upgrade (including a Postgres upgrade) seems to have sorted the issue. Though in our case it is two separate environment (configured the same way automatically) we did not upgraded as per say...
>
> Now I have a network issue on Contacts.app when syncing ~5000 contacts!
> It is never ending... getting there... getting there... :-)
>
> Just though to keep you posted in case it may be useful to others.
>
> Cheers,
>
>> On Wed, May 4, 2016 at 11:04 AM, Richard LEGER <richard.leger@gmail.com> wrote:
>>
>> Marten,
>>
>> I am encountering similar database deadlock issue as Egoitz described and try to fix (here https://sourceforge.net/p/davical/mailman/message/34986381/ and here https://sourceforge.net/p/davical/mailman/message/35006191/) with Postgres on Linux.
>>
>> For testing with client, I used various version of Contacts.app on Mac OS X 10.9, 10.10, 10.11...
>>
>> With the following config, no issue...
>> When I try to sync 5 contacts, five are effectively stored to the server... the others are not...
>>
>> Srl config:
>>
>> AWL version 0.55
>> Davical version 1.1.3.1
>> psql (PostgreSQL) 9.3.12
>> PHP 5.5.9-1ubuntu4.16 (cli) (built: Apr 20 2016 14:31:27)
>> Copyright (c) 1997-2014 The PHP Group
>> Zend Engine v2.5.0, Copyright (c) 1998-2014 Zend Technologies
>> with Zend OPcache v7.0.3, Copyright (c) 1999-2014, by Zend Technologies
>> Distributor ID: Ubuntu
>> Description: Ubuntu 14.04.4 LTS
>> Release: 14.04
>> Codename: trusty
>> Server version: Apache/2.4.7 (Ubuntu)
>> Server built: Jan 14 2016 17:45:23
>>
>> With the following config... the errors like the one further below appear...
>> When I try to sync 5 contacts, only one is stored to server... the others are not...
>>
>> Wood config:
>>
>> AWL version 0.56
>> Davical version 1.1.4
>> psql (PostgreSQL) 9.1.21
>> contains support for command-line editing
>> PHP 5.3.10-1ubuntu3.22 with Suhosin-Patch (cli) (built: Apr 20 2016 12:31:15)
>> Copyright (c) 1997-2012 The PHP Group
>> Zend Engine v2.3.0, Copyright (c) 1998-2012 Zend Technologies
>> Distributor ID: Ubuntu
>> Description: Ubuntu 12.04.5 LTS
>> Release: 12.04
>> Codename: precise
>> Server version: Apache/2.2.22 (Ubuntu)
>> Server built: Jul 24 2015 17:25:42
>>
>>> Postgres error:
>>> "2016-05-04 07:42:19 BST ERROR: deadlock detected
>>> 2016-05-04 07:42:19 BST DETAIL: Process 6162 waits for ShareLock on
>>> transaction 192145; blocked by process 6164.
>>> Process 6164 waits for ExclusiveLock on tuple (1,27) of relation
>>> 17248 of database 17112; blocked by process 6162.
>>> Process 6162: INSERT INTO caldav_data ( user_no, dav_name,
>> dav_etag,
>>> caldav_data, caldav_type, logged_user, created, modified,
>> collection_id )
>>> VALUES( 1006,
>>> '/richard/addresses/ac2fee95-ef58-451d-ae1d-58c3f6eca1b6.vcf',
>>> '61413f34312ce41687ee8e4fc1a22c1f', E'BEGIN\x3aVCARD
>>> (...)
>>> "
>>
>> That sounds likes there are 2 Apache processes (6162 and 6164) both trying to access the same table ("caldav_data"). Process 6164 is waiting for process 6162 to finish.
>>
>> Issue raise with both PUT and DELETE commands...
>>
>> No error reported on client during PUT sync...
>> HTTP 500 error reported on client which stop to sync during DELETE...
>>
>> As in the first config case the issue does not raise, we believe it might be fixable in Davical...
>>
>> A short term fix we are trying is limiting the Apache connections for Davical to only one per IP address (simplest but
>> not best solution), a Davical implemented solution may be best...
>>
>> Any advice how to sort this issue?
>
> ------------------------------------------------------------------------------
> Find and fix application performance issues faster with Applications Manager
> Applications Manager provides deep performance insights into multiple tiers of
> your business applications. It resolves application problems quickly and
> reduces your MTTR. Get your free trial!
> https://ad.doubleclick.net/ddm/clk/302982198;130105516;z
> _______________________________________________
> Davical-general mailing list
> Davical-general@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/davical-general
Re: [DAViCal-devel] Possible fix for deadlocks in database [ In reply to ]
Hi!

Although should say I used Davical 1.1.4 for applying the patch

Regards,

El 5/5/16 a las 1:24, Egoitz Aurrekoetxea escribió:
> Hi!!
>
> For the deadlock issue you could use my patch. I use it in production
> without issues and fixes that problems.
>
> Best regards,
>
> Egoirz Aurrekoetxea Aurre
> egoitz@ramattack.net <mailto:egoitz@ramattack.net>
> Sent from my smartphone
>
> El 4 may 2016, a las 19:45, Richard LEGER <richard.leger@gmail.com
> <mailto:richard.leger@gmail.com>> escribió:
>
>> Hi,
>>
>> Just to give you an update...
>>
>> Upgrading Davical on our Srl config environment...
>>
>> from:
>> AWL version 0.55
>> Davical version 1.1.3.1
>>
>> to:
>> AWL version 0.56
>> Davical version 1.1.4
>>
>> ... confirmed that the issue is not linked to Davical but to Postgres
>> or else...
>>
>> Indeed in both configuration in our Srl config environment we obtain
>> the same result, bulk PUT and DELETE complete successfully without
>> Postgres errors (deadlock) as previously encountered in our Woody
>> config environment.
>>
>> So a Distro upgrade (including a Postgres upgrade) seems to have
>> sorted the issue. Though in our case it is two separate environment
>> (configured the same way automatically) we did not upgraded as per say...
>>
>> Now I have a network issue on Contacts.app when syncing ~5000 contacts!
>> It is never ending... getting there... getting there... :-)
>>
>> Just though to keep you posted in case it may be useful to others.
>>
>> Cheers,
>>
>> On Wed, May 4, 2016 at 11:04 AM, Richard LEGER
>> <richard.leger@gmail.com <mailto:richard.leger@gmail.com>> wrote:
>>
>>
>> Marten,
>>
>> I am encountering similar database deadlock issue as Egoitz
>> described and try to fix (here
>> https://sourceforge.net/p/davical/mailman/message/34986381/ and
>> here https://sourceforge.net/p/davical/mailman/message/35006191/)
>> with Postgres on Linux.
>>
>> For testing with client, I used various version of Contacts.app
>> on Mac OS X 10.9, 10.10, 10.11...
>>
>> With the following config, no issue...
>> When I try to sync 5 contacts, five are effectively stored to the
>> server... the others are not...
>>
>> Srl config:
>>
>> AWL version 0.55
>> Davical version 1.1.3.1
>> psql (PostgreSQL) 9.3.12
>> PHP 5.5.9-1ubuntu4.16 (cli) (built: Apr 20 2016 14:31:27)
>> Copyright (c) 1997-2014 The PHP Group
>> Zend Engine v2.5.0, Copyright (c) 1998-2014 Zend Technologies
>> with Zend OPcache v7.0.3, Copyright (c) 1999-2014, by Zend
>> Technologies
>> Distributor ID: Ubuntu
>> Description: Ubuntu 14.04.4 LTS
>> Release: 14.04
>> Codename: trusty
>> Server version: Apache/2.4.7 (Ubuntu)
>> Server built: Jan 14 2016 17:45:23
>>
>> With the following config... the errors like the one further
>> below appear...
>> When I try to sync 5 contacts, only one is stored to server...
>> the others are not...
>>
>> Wood config:
>>
>> AWL version 0.56
>> Davical version 1.1.4
>> psql (PostgreSQL) 9.1.21
>> contains support for command-line editing
>> PHP 5.3.10-1ubuntu3.22 with Suhosin-Patch (cli) (built: Apr 20
>> 2016 12:31:15)
>> Copyright (c) 1997-2012 The PHP Group
>> Zend Engine v2.3.0, Copyright (c) 1998-2012 Zend Technologies
>> Distributor ID: Ubuntu
>> Description: Ubuntu 12.04.5 LTS
>> Release: 12.04
>> Codename: precise
>> Server version: Apache/2.2.22 (Ubuntu)
>> Server built: Jul 24 2015 17:25:42
>>
>>> Postgres error:
>>> "2016-05-04 07:42:19 BST ERROR: deadlock detected
>>> 2016-05-04 07:42:19 BST DETAIL: Process 6162 waits for ShareLock on
>>> transaction 192145; blocked by process 6164.
>>> Process 6164 waits for ExclusiveLock on tuple (1,27) of relation
>>> 17248 of database 17112; blocked by process 6162.
>>> Process 6162: INSERT INTO caldav_data ( user_no, dav_name,
>>
>> dav_etag,
>>
>>> caldav_data, caldav_type, logged_user, created, modified,
>>
>> collection_id )
>>
>>> VALUES( 1006,
>>> '/richard/addresses/ac2fee95-ef58-451d-ae1d-58c3f6eca1b6.vcf',
>>> '61413f34312ce41687ee8e4fc1a22c1f', E'BEGIN\x3aVCARD (...) "
>> That sounds likes there are 2 Apache processes (6162 and 6164)
>> both trying to access the same table ("caldav_data"). Process
>> 6164 is waiting for process 6162 to finish.
>>
>> Issue raise with both PUT and DELETE commands...
>>
>> No error reported on client during PUT sync...
>> HTTP 500 error reported on client which stop to sync during DELETE...
>>
>> As in the first config case the issue does not raise, we believe
>> it might be fixable in Davical...
>>
>> A short term fix we are trying is limiting the Apache connections
>> for Davical to only one per IP address (simplest but
>> not best solution), a Davical implemented solution may be best...
>>
>> Any advice how to sort this issue?
>>
>>
>> ------------------------------------------------------------------------------
>> Find and fix application performance issues faster with Applications
>> Manager
>> Applications Manager provides deep performance insights into multiple
>> tiers of
>> your business applications. It resolves application problems quickly and
>> reduces your MTTR. Get your free trial!
>> https://ad.doubleclick.net/ddm/clk/302982198;130105516;z
>> _______________________________________________
>> Davical-general mailing list
>> Davical-general@lists.sourceforge.net
>> <mailto:Davical-general@lists.sourceforge.net>
>> https://lists.sourceforge.net/lists/listinfo/davical-general
>
>
> ------------------------------------------------------------------------------
> Find and fix application performance issues faster with Applications Manager
> Applications Manager provides deep performance insights into multiple tiers of
> your business applications. It resolves application problems quickly and
> reduces your MTTR. Get your free trial!
> https://ad.doubleclick.net/ddm/clk/302982198;130105516;z
>
>
> _______________________________________________
> Davical-general mailing list
> Davical-general@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/davical-general

--


sarenet
*Egoitz Aurrekoetxea*
Departamento de sistemas
944 209 470
Parque Tecnológico. Edificio 103
48170 Zamudio (Bizkaia)
egoitz@sarenet.es <mailto:egoitz@sarenet.es>
www.sarenet.es <http://www.sarenet.es>

Antes de imprimir este correo electrónico piense si es necesario hacerlo.
Re: [DAViCal-devel] Possible fix for deadlocks in database [ In reply to ]
On 4 May 2016 18:45:24 BST, Richard LEGER <richard.leger@gmail.com> wrote:
>
>Now I have a network issue on Contacts.app when syncing ~5000 contacts!

Excellent news!

The network issue encountered "The operation couldn't be completed. (CoreDAVErrorDomain error 1.)" seemed to have cleared itself out automatically. Just left Contacts.app opened for a while and following further auto-sync issue cleared... Adressbook is all up-to-date now and seems to work as expected.

Apparently this issue is also encountered by iCloud users so may well be related to the Contacts.app in Mac OS X 10.11.

No issue encountered with Contacts.app client on Mac OS X 10.10. Don't have 10.9 to test with ;-)

Thanks for the tip Egoirz but I can confirm your patch is not necessarily required with Davical 1.1.4 on Ubuntu 14.04.

It may just be a matter for you to update some packages on your environment (any info on your environment?) starting with the database one if possible.

Long term, it is always better to rely on DB locking system anyway... than "manual" locking via files within Davical ;-)

Cheers,
Richard


------------------------------------------------------------------------------
Find and fix application performance issues faster with Applications Manager
Applications Manager provides deep performance insights into multiple tiers of
your business applications. It resolves application problems quickly and
reduces your MTTR. Get your free trial!
https://ad.doubleclick.net/ddm/clk/302982198;130105516;z
_______________________________________________
Davical-general mailing list
Davical-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/davical-general
Re: [DAViCal-devel] Possible fix for deadlocks in database [ In reply to ]
Good morning,

Sure it is, and allows scalling to several servers and so.... but I
needed a fast, functional and at least temporal solution... the
difference between Contacts in 10,9, 10.10 and 10.11 is precisely this.
Contact app does multiple uploads
at the same time in 10.11 but not in 10.9 or 10.10 where it does one by
one. It will probably do the same way with iCloud.... and so.. .the same
issue probably.... perhaps they used some Davical modified version....

When that message appears in contacts (CoreDAVErrorDomain) it dissapears
only and keeps later syncing.... but have seen too sometimes doing weird
things with contacts, as duplications or contact dissapearings after
this CoreDAVErrorDomain
in 10.11 when suffering this issue...

After applying my patch all worked smoothly... else... apart from not
uploading all at first time.... some weird things happened sometimes....

Bye!!


El 5/5/16 a las 11:26, Richard Leger escribió:
>
> On 4 May 2016 18:45:24 BST, Richard LEGER <richard.leger@gmail.com> wrote:
>> Now I have a network issue on Contacts.app when syncing ~5000 contacts!
> Excellent news!
>
> The network issue encountered "The operation couldn't be completed. (CoreDAVErrorDomain error 1.)" seemed to have cleared itself out automatically. Just left Contacts.app opened for a while and following further auto-sync issue cleared... Adressbook is all up-to-date now and seems to work as expected.
>
> Apparently this issue is also encountered by iCloud users so may well be related to the Contacts.app in Mac OS X 10.11.
>
> No issue encountered with Contacts.app client on Mac OS X 10.10. Don't have 10.9 to test with ;-)
>
> Thanks for the tip Egoirz but I can confirm your patch is not necessarily required with Davical 1.1.4 on Ubuntu 14.04.
>
> It may just be a matter for you to update some packages on your environment (any info on your environment?) starting with the database one if possible.
>
> Long term, it is always better to rely on DB locking system anyway... than "manual" locking via files within Davical ;-)
>
> Cheers,
> Richard
>
>
> ------------------------------------------------------------------------------
> Find and fix application performance issues faster with Applications Manager
> Applications Manager provides deep performance insights into multiple tiers of
> your business applications. It resolves application problems quickly and
> reduces your MTTR. Get your free trial!
> https://ad.doubleclick.net/ddm/clk/302982198;130105516;z
> _______________________________________________
> Davical-general mailing list
> Davical-general@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/davical-general

--


sarenet
*Egoitz Aurrekoetxea*
Departamento de sistemas
944 209 470
Parque Tecnológico. Edificio 103
48170 Zamudio (Bizkaia)
egoitz@sarenet.es <mailto:egoitz@sarenet.es>
www.sarenet.es <http://www.sarenet.es>

Antes de imprimir este correo electrónico piense si es necesario hacerlo.
Re: [DAViCal-devel] Possible fix for deadlocks in database [ In reply to ]
On 5 May 2016 10:26:30 BST, Richard Leger <richard.leger@gmail.com> wrote:

>It may just be a matter for you to update some packages on your
>environment (any info on your environment?) starting with the database
>one if possible.

Sorry by environement I meant:

OS version
AWL version
Database version
PHP version
Apache version

You already confirmed you use Davical 1.1.4 ;-)


------------------------------------------------------------------------------
Find and fix application performance issues faster with Applications Manager
Applications Manager provides deep performance insights into multiple tiers of
your business applications. It resolves application problems quickly and
reduces your MTTR. Get your free trial!
https://ad.doubleclick.net/ddm/clk/302982198;130105516;z
_______________________________________________
Davical-general mailing list
Davical-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/davical-general
Re: [DAViCal-devel] Possible fix for deadlocks in database [ In reply to ]
Hi!

Don't really know if just that is meaningful for not having the issue
but... anyway... let us know how all evolutes.... and if making changes
and so.... is all fine :)

I run FreeBSD 10-release, apache 2.4 php 5.4 (php mod) and postgres
9.0... the rest the last versions....

but have tested it in Postgres 9.4 and same history.... I think the env
is not... totally meaningful.... in this issue.... are you sure you
can't now reproduce the issue? or that you have
found a workaround?, how did you finally do?.... IMHO is not a way of
solving just letting it to return CoreDAVErrorDomain from time to time
till it syncs.... which is what I have understood
about your around... am I wrong?

Best regards,

El 5/5/16 a las 12:37, Richard Leger escribió:
>
> On 5 May 2016 10:26:30 BST, Richard Leger <richard.leger@gmail.com> wrote:
>
>> It may just be a matter for you to update some packages on your
>> environment (any info on your environment?) starting with the database
>> one if possible.
> Sorry by environement I meant:
>
> OS version
> AWL version
> Database version
> PHP version
> Apache version
>
> You already confirmed you use Davical 1.1.4 ;-)
>
>
> ------------------------------------------------------------------------------
> Find and fix application performance issues faster with Applications Manager
> Applications Manager provides deep performance insights into multiple tiers of
> your business applications. It resolves application problems quickly and
> reduces your MTTR. Get your free trial!
> https://ad.doubleclick.net/ddm/clk/302982198;130105516;z
> _______________________________________________
> Davical-general mailing list
> Davical-general@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/davical-general

--


sarenet
*Egoitz Aurrekoetxea*
Departamento de sistemas
944 209 470
Parque Tecnológico. Edificio 103
48170 Zamudio (Bizkaia)
egoitz@sarenet.es <mailto:egoitz@sarenet.es>
www.sarenet.es <http://www.sarenet.es>

Antes de imprimir este correo electrónico piense si es necesario hacerlo.
Re: [DAViCal-devel] Possible fix for deadlocks in database [ In reply to ]
Re: [Davical-general] [DAViCal-devel] Possible fix for deadlocks in database
From: Egoitz Aurrekoetxea <egoitz@sa...> - 2016-05-05 11:34:42

> Don't really know if just that is meaningful for not having the issue
but... anyway...

Providing info on environment may help others having the same issue... Also it may help identify what is common or not on platformd that encounter the same issues... and narrow down the possible cause...

>let us know how all evolutes.... and if making changes and so.... is all fine :)

We will put in production at client soon so we would know if further issue raise via real time life experience ;-)
Keep posted.

> I run FreeBSD 10-release, apache 2.4 php 5.4 (php mod) and postgres
9.0... the rest the last versions....

Thanks for letting us know ;-)

> but have tested it in Postgres 9.4 and same history.... I think the env
is not... totally meaningful.... in this issue....

Well for us with a complete different version of environment (setup the exact same way) both with Davical 1.1.4... does make a difference so it is clearly an "enviromental" issue...

> are you sure you can't now reproduce the issue?

Yes. No more DB deadlock in our Prod environment (SRL), still deadlock in our Test environment (Woody) with older OS realease and packages versions...

> or that you have found a workaround?how did you finally do?....

Once again, no workaround applied, just changed environment on which Davical 1.1.4 runs (with updates OS realease and package versions including Postgres).

>IMHO is not a way of
solving just letting it to return CoreDAVErrorDomain from time to time
till it syncs.... which is what I have understood
about your around... am I wrong?

In my case error cleared itself after few autosync and did not reappear... I believe error occurs because client conn timeout prior completion of the sync... then it resume...

On Thu, May 5, 2016 at 11:01 AM, <
davical-general-request@lists.sourceforge.net> wrote:
>
>
> Date: Thu, 5 May 2016 12:01:18 +0200
> From: Egoitz Aurrekoetxea <egoitz@sarenet.es>
> Subject: Re: [Davical-general] [DAViCal-devel] Possible fix for deadlocks in database
>
> Sure it is, and allows scalling to several servers and so.... but I
> needed a fast, functional and at least temporal solution...

Quick fix works when you need it. Thanks for sharing your finding. I do not exclude it may be useful to us at some point ;-)

> the
> difference between Contacts in 10,9, 10.10 and 10.11 is precisely this.
> Contact app does multiple uploads
>

An alternative could be to limit within Apache to one connection max per IP for Davical... but haven't tested yet... as current solutiin works for us...

> When that message appears in contacts (CoreDAVErrorDomain) it dissapears
> only and keeps later syncing.... but have seen too sometimes doing weird
> things with contacts, as duplications or contact dissapearings after
> this CoreDAVErrorDomain
> in 10.11 when suffering this issue...

Thanks for letting me know. I haven't tried your patch just yet as it it is currently not necessary but I keep it in mind as a possible quick fix in the eventuality of issues you describe above would occur.

> After applying my patch all worked smoothly... else... apart from not
> uploading all at first time.... some weird things happened sometimes....

I agree ;-)
In our case the first import of ~5000 contacts over Internet worked like a charmed and took less than a minute (ADSL client side, FTTC server side).