Grommunio 2025.1.2 Webmail is extremly slow

gbailey · 2025-04-24T11:54:18+00:00

Hy, i have Grommunio Appliance installed, and since 1 Month, or with 30 Clients, the Webclient is very slow. The Appliance run on Proxmox as KVM with 30GB R...

WalterH

crpb or sqlite

do you have a SQLite query to list the search folders?

crpb

WalterH do you have a SQLite query to list the search folders?

Yeah, they will also show up with the simple temp.folderlist

-- FOLDER LOOKUP VIEW VIA FOLDERS INCLUDING PARENT
-- DROP VIEW IF EXISTS temp.folderlist;
CREATE VIEW IF NOT EXISTS temp.folderlist
AS SELECT distinct(p.folder_id),
          p.parent_id,
          f1.propval AS foldername
     FROM folders p
     JOIN folder_properties f1
       ON f1.folder_id = p.folder_id
      AND f1.proptag = 805371935
    ORDER BY p.folder_id
;
SELECT  * FROM temp.folderlist WHERE lower(foldername) LIKE '%search%' ;

WalterH

gbailey If the system is only now updating to EV-18, your system was not patched to the latest version when you carried out the ICS repair. The ICS repair cannot find anything, the additions for the ICS repair have only been in grommunio for a few days.

crpb

Wobei du auch direkt auf folders.search_flags > 0 nen query machen kannst so wie ich das hier gerade sehe und vergleiche

gbailey

@WalterH thank's for the hint.
after that, the repair script takes 3 Minutes an repair 60k messages. 👍️

@crpb ich dachte auf sqlite darf immer nur 1 Process zugreifen, aber anscheind kann man im Betrieb echt die DB abfragen.
So langsam verstehe ich, warum Sie beim Fork auf sqlite gesetzt haben. 😃

crpb

gbailey @crpb ich dachte auf sqlite darf immer nur 1 Process zugreifen, aber anscheind kann man im Betrieb echt die DB abfragen.

Ich lasse alle diese sachen außerdem immer mit sqlite3 -readonly laufen um ganz sicher zu sein. Also den selector.

Aber ja, dadurch das sqlite3 wal benutzt können da auch mehrere lesen und auch welche schreiben.

WalterH

gbailey after that, the repair script takes 3 Minutes an repair 60k messages.

Is the speed issue now solved?

gbailey

WalterH: After a week, the server is running smoothly again.

So the problem is solved.

Thanks everyone for your help.

StefanRi25

Hi good morning ... maybe you can tell me what happended ... I installed the 2025.01.2 OVA for VMWare last Sunday ... for one week everything was working fine and I was happy. Yesterday evening it started that all 3 Users (ok 2 with OST file) on the Appliance lost connection to the Server every 2 or 3 minutes (that the reason why I started searching and am here now) ... I was confused ... removed the Autodiscover SRV Record - as it was mentioned - but that didn't helped. After another meanwhile my test emails did not appear in the mailboxes (whether Outlook nor through the Webinterface) any more while in the SMTP Log it was visible that the should have been delivered. The user where LDAP synced from Active directory. I'm now running with the backup from last Friday ...if you need any further infos please tell me ... the VM is still there and can be started ... any help would be appreciated ...

TylerDurden

StefanRi25 If you are on the community edition - the current Groxmox Edition has a bug and is segfaulting once a Outlook user connects. Everything is "fine" as long as not Outlook connects.
With outlook connections, it crashed every 2 minutes. So far no update ....

StefanRi25

Hi Tyler, thanks for your quick answer, but I think here I’m responsible for the problem.
When I installed gromox first I thought it is a fine to have an Active Directory sync and so I synced 3 users one group and a contact … a little later that evening I thought this isn’t such a good idea for beginning - so I cut off the AD sync and here as it seems something went wrong.
The users were cut off but the AD group remained somewhere in the background and caused many many errors with the Gromox zcore - HX_inet_connect to [::1]:5000 – which makes access with Outlook very slow especially when using in online mode without an OST / local cache.
Additional later I maybe damaged some mailboxes because of using the Outook‘s cache mode not only for the primary mailbox.
Yesterday in the evening I switched to an backup and noticed this. I started again with AD sync and realized the remaining AD group and that Outook access becomes slow and also looses it’s connection to the Gromox Server the more often the more errorr of - HX_inet_connect to [::1]:5000 - there are.
I now using TLS to communicate to the Domain Controller – I didin’t managed to get LAPS completly working until now (Firewall) … I suppose when this is working no delay will be there any more.
At the moment Gromox runs fine and until now I realized no problems when only using the primary mailbox with cache … I hope I will get LAPS working the next days
I really like Gromox … when I started searching for an Exchange replacement I was fascinated - yeeeeaaaahhh that’s exatly what I was searching for …
(and: Don’t know if you have done an Exchange onprem update the last time … that’s absolutely no fun …. they force you to switch to Kerberos when updating to Exchange SE …. many many fun 😉)

Thanks and have a nice evening

TylerDurden

StefanRi25
If it should nevertheless have been the mentioned bug - the fix is just out in the community repository. 🙂

PS: " HX_inet_connect to [::1]:5000" would have been a sign for that bug.

StefanRi25

Hi again .... wooow ... just installed the update and for me the " HX_inet_connect to [::1]:5000" error seems to be gone ... and Outlook is as fast as it is with Exchange even without cache mode ... very cool ... thanks for the info :-)

« Previous Page