Issue after upgrading to 8.8.15

Ask questions about your setup or get help installing ZCS server (ZD section below).
hsingh
Zimbra Employee
Zimbra Employee
Posts: 5
Joined: Tue Aug 02, 2016 1:40 pm

Re: Issue after upgrading to 8.8.15

Post by hsingh »

Hello All,

We have a workaround for this, please enable zimbraCsrfTokenCheckEnabled by running below commands and restart mailbox service:

Code: Select all

su - zimbra 
zmprov mcf zimbraCsrfTokenCheckEnabled TRUE
zmmailboxdctl restart
User avatar
benoit.georgelin
Posts: 4
Joined: Mon Jun 09, 2014 10:27 am

Re: Issue after upgrading to 8.8.15

Post by benoit.georgelin »

Hello,

In our situation, this fix does not help at all.
We are still having a lot of :

Code: Select all

2019-11-20 13:30:13,655 WARN  [qtp1708169732-28478:https://xxxxxx/zimbra/m/zmain] [] webclient - system failure: Connection pool shut down
After restarting Zimbra it does not even last 24h
Is there any way to investigate this more ?

We are in multi-mode server with store-ui and proxy, maybe that's the difference ?

Thanks for your help,
Regards
User avatar
L. Mark Stone
Ambassador
Ambassador
Posts: 2796
Joined: Wed Oct 09, 2013 11:35 am
Location: Portland, Maine, US
ZCS/ZD Version: 10.0.6 Network Edition
Contact:

Re: Issue after upgrading to 8.8.15

Post by L. Mark Stone »

benoit.georgelin wrote:Hello,

In our situation, this fix does not help at all.
We are still having a lot of :

Code: Select all

2019-11-20 13:30:13,655 WARN  [qtp1708169732-28478:https://xxxxxx/zimbra/m/zmain] [] webclient - system failure: Connection pool shut down
After restarting Zimbra it does not even last 24h
Is there any way to investigate this more ?

We are in multi-mode server with store-ui and proxy, maybe that's the difference ?

Thanks for your help,
Regards
We and one of our customers had this issue. In our cases, the symptoms were that web client sessions generated a "Server Error 500" for the users, but that IMAP and other non web-ui access methods continued to function.

Zimbra Support have identified the root cause and given us a bug number.

If your symptoms are identical to ours, the interim workaround is primarily to identify all mailboxes that poll external accounts, and increase the polling interval to four hours or more. The bug is part of Java, where Java httpclient connections are not released quickly enough, Zimbra gets stuck in an infinite loop, eventually you run out of connections and the entire connection pool shuts down. The Java bug report is here: https://bugs.openjdk.java.net/browse/JDK-8214129

After you identify the mailboxes which are polling external accounts, and you have the data source names, you can adjust the data source polling interval, like so for example:

Code: Select all

zmprov mds user@yourdomain.com "Gmail" zimbraDataSourcePollingInterval 4h
To get all the data sources configured on a mailbox (and the name which you will need to use in the above command), you can run:

Code: Select all

zmprov mds user@yourdomain.com | grep " name "
Hope that helps,
Mark
___________________________________
L. Mark Stone
Mission Critical Email - Zimbra VAR/BSP/Training Partner https://www.missioncriticalemail.com/
AWS Certified Solutions Architect-Associate
User avatar
benoit.georgelin
Posts: 4
Joined: Mon Jun 09, 2014 10:27 am

Re: Issue after upgrading to 8.8.15

Post by benoit.georgelin »

Hi Mark.

I do appreciate a lot your answer.
It's exactly the issue that we are facing..

We run a lof of mailboxes and a lot of users uses external accounts I guess.
We will have to figure how to change the polling for all those external accounts. I think we can user the classes service to set the time globaly , is that right ?

Thanks
User avatar
L. Mark Stone
Ambassador
Ambassador
Posts: 2796
Joined: Wed Oct 09, 2013 11:35 am
Location: Portland, Maine, US
ZCS/ZD Version: 10.0.6 Network Edition
Contact:

Re: Issue after upgrading to 8.8.15

Post by L. Mark Stone »

benoit.georgelin wrote:Hi Mark.

I do appreciate a lot your answer.
It's exactly the issue that we are facing..

We run a lof of mailboxes and a lot of users uses external accounts I guess.
We will have to figure how to change the polling for all those external accounts. I think we can user the classes service to set the time globaly , is that right ?

Thanks
You can indeed set the minimum polling interval via COS; I have not tested however how inheritance works for this variable. IOW, if a user has an external data source already configured, the minimum polling interval may already be set for that data source.

If so, then it would work like the Server Pool restrictions in a multi-mailstore environment: you can restrict new mailbox creation to specific mail stores, but making that change doesn't move existing mailboxes on other servers to the newly designated ones.

Certainly Zimbra Support's instructions to me were to modify the polling interval at the account-data source level, not at the Class of Service level.

If you test this (create a data source for an account with a 5-minute polling interval, then change the class of service to a 1 hour polling interval, and then see if the polling interval already set for that data source is increased), please post the results to benefit others until this bug is fixed?

Hope that helps,
Mark
___________________________________
L. Mark Stone
Mission Critical Email - Zimbra VAR/BSP/Training Partner https://www.missioncriticalemail.com/
AWS Certified Solutions Architect-Associate
sanjitsahu
Posts: 3
Joined: Tue May 28, 2019 6:42 am

Re: Issue after upgrading to 8.8.15

Post by sanjitsahu »

I have upgraded my ZCS from 8.8.12 to 8.8.15 after which I am unable to login through web client which shows HTTP 502 Error . zmprov command is also showing error "ERROR: service.PROXY_ERROR (error while proxying request to target server: Service Unavailable)" .
zmcontrol -v :Release 8.8.15_GA_3869.RHEL6_64_20190917004220 RHEL6_64 FOSS edition, Patch 8.8.15_P5.

The upgradation is done smoothly but even after restarting the service as well as server , problem is still there. Please be kind enough to help me out.

Thanks
Sanjit

Edit - 1:
By deleting cookies / history etc I could see login page but unable to login and the error is "A Network service error has occurred" . zmprov still shows above error
manuel.brignoli
Posts: 4
Joined: Thu Aug 04, 2016 10:35 am

Re: Issue after upgrading to 8.8.15

Post by manuel.brignoli »

Hi,

Same here... After applying Patch P5 on ZCS 8.8.15 on Centos 6 and Centos 7.
alladin_bp
Posts: 3
Joined: Wed Dec 18, 2019 10:55 am

Re: Issue after upgrading to 8.8.15

Post by alladin_bp »

Please see the other topic on the recent patch v5 problems: viewtopic.php?f=13&t=67491
sanjitsahu
Posts: 3
Joined: Tue May 28, 2019 6:42 am

Re: Issue after upgrading to 8.8.15

Post by sanjitsahu »

Thanks alladin_bp , Issue solved after removing /opt/zimbra/lib/ext/zimbradrive/zal.jar . I also see a bug in Admin console. If one go to admin console through the link given in user home screen, then the port it is redirecting is 9091 instead of 7071. e.g https://mail.xxxx.xxx:9071/zimbraAdmin which should be https://mail.xxxx.xxx:7071/zimbraAdmin
lfasci
Advanced member
Advanced member
Posts: 145
Joined: Mon Dec 09, 2013 2:38 am

Re: Issue after upgrading to 8.8.15

Post by lfasci »

Find all zal.yar

find /opt/zimbra/lib -name zal.jar

The zal.jar file could be located in /opt/zimbra/lib/ext/zimbradrive if you installed the owncloud zimlet.

If you don't use drive or ownclod zimlet you can

Remove te zimbra-drive package using yum or apt
Remove the zimlet using zmziletctl undeploy zimletName

After this, please check the /opt/zimbra/lib/ext/ subfolder find /opt/zimbra/lib -name zal.jar
and remove all zal.jar except /opt/zimbra/lib/ext/network-modules-ng/zal.jar!

Empty the /opt/zimbra/jetty/work/ folder

restart the mailbox

su - zimbra
zmmailboxdctl restart

If you use the zimlet contact the developer for furter info, or just try to make a symlink to the file /opt/zimbra/lib/ext/network-modules-ng/zal.jar

Bye
Lorenzo
Post Reply