scalix 11.0.3 installation borked & scalix 11.0.2.1 work

[daniel.sand]

Greetings,

i evaluated scalix 11.0.2.1 one month ago, and it worked like a charm. Installation quick and uncomplicated out of the box. I was really impressed.

so now i was going on step further on testing this fine software.
I just spend one day on 11.0.3 installation. It was really strange.

The installation self, no problems.

But after i was trying to send emails to 2 users in the system over webmail / Outlook nothing happend. the mail didnt get delivered.

so i tried to get it working, looked into sendmail what was happening and checked the base installation. sendmail told me that the user didnt exist and rejected the mail. So i looked into the the scalix system with omshowu if this was correct and it wasn't correct the user was there.

I checked the ldapmapper and turend on the debug the requests... allright he was checking his user@hostname and the user@domain i choosed, but no user was found.
i spend the whole day in deinstalling reinstalling and bug searching but my sendmail kung-fu wasnt good enough ( i'm more the postfix user ) and my scalix level is below basic.

so i deceided to go back to 11.0.2.1... and guess what happend. All worked out of the box...

for my test i will work further with 11.0.2.1 but maybe someone should look into the 11.0.3 release.

before i forget, installed on:

Fedora Core release 5

latest updates installed.

[jaime.pinto]

Hi Daniel

Indeed release 11.0.3 of the server is a bit broken (in very obvious places), despite all the other fixes, and the guys at Scalix are aware of that. There are several references to that on this forum.

The best combination yet of the server/connector combo was 11.0.2/11.0.3. For anybody out there that has not done the upgrade yet, if you have the choice at this point just skip 11.0.3 of the server all together. You'll be better off.

On the other hand you can just hand tight. The 11.0.4 combo should be just around the corner.

Jaime

[daniel.sand]

ah k.

thx for the info, maybe i used the wrong search terms, but i only find problems on updateing this versions, not on fresh installs.

my force was only to offer some information and my feedback to offer some kind of QoS :]

so nevermind, ah the big BUT !

it would be a good idear to remove the sources to this borked version. It will only cause big disturbances...

[Shredder]

I think you can solve problem by putting your mailserver FQDN and domainname in /etc/mail/local-host-names

ie.
mail.domain.com
domain.com

That seems to fix those kind of issues.

Shredder

[daniel.sand]

I think you can solve problem by putting your mailserver FQDN and domainname in /etc/mail/local-host-names
ie.
mail.domain.com
domain.com
That seems to fix those kind of issues.

not really. you get the "dog bites his tail"-effect.

sendmail now thinks its for sure the owner for this passes this to the ldapmapper and the ldapmapper still say _no_ after that im not quiet sure what happend it was a dead-lock or it delivered the mail locally via procmail fallback.

i was one step further and edited smtp.cfg of scalix added the local domain but the effect was for sure not the right one.

[florian]

Jaime,

I've been monitoring your posts for a while and in general you show good judgement I'd say; this one surprises me a bit, though.

What are your specific claims that 11.0.3 is broken over 11.0.2 on the server side based on, in particular in the context that's talked about here. Given the list of server bugs fixed and the list of changes done, I simply find it hard to see any evidence in connection with this thread.

Looking forward,
Florian.

[jaime.pinto]

Hi Florian

Simply put, after the 11.0.3 server upgrade, about 2-3 weeks ago, the whole system became worst for our requirements. Bottom like, once we reached a certain plateau, going backwards generated a lot of disappointment, and an avalanche of complaints. So much that some on the management vented out their frustrations, and wanted to pull the plug on the whole scalix exercise!

Prior to that we had finally reached a certain level of stability with 11.0.3 of the connector and 11.0.2 of the server. It was not all we hoped for, but it was much better than the difficulties we had experienced since mid January. If it wasn't for the proximity of the 11.0.4 release I'd step back to 11.0.2.1 of the server. It so happens that it took several days for us to come to the realization it was not misuse on the part of the users.

The upgrade was done by our reseller, but I followed his steps closely, and there was nothing odd about anything. He had already done several other upgrades prior to doing ours. So, unless some very important step was missed during the upgrade, I have nothing else to blame but the release itself. Maybe a fresh installation (new server, new accounts, no upgrades) would have rendered a different outcome. Regardless, the net effect was dissatisfaction

If I could summarize what I'd hope for *now* with one word it would be *stability*. By that I mean *no bugs*:
1) being able to utilize Scalix without emails and sometimes whole folders disappearing from OL2003, SWA and even IMAP, as it happened after the 11.0.3 upgrade of the server.

2) This SmartCache flip/flop mess also needs a solution. We would like to see the connector 100% problem free when SmartCache is OFF, and to be able to sync the BlackBerries with the USB cable without the choking of calendar/contact entries anymore (it's happening again!).

3) No more profile recreation. We should be able to switch SmartCache ON/OFF on the fly, even after the installation. SmartCache is not imperative for us at the moment, but we would like to take advantage of improved speed. Should we detect a hint of a problem that could be attributed to SmartCache we can not subject the end user to the intrusion and incovenience of having to delete and create a profile again.

4) I better developed *upgrade script* (NOT INSTALL), that takes into consideration that a server has been in place from before, and "respects" the customization done previously.

I can think of a number of other items I'd consider enhancement requests or wish list, but they can wait until 11.0.5 or later. I see no point distracting from the above immediate priorities at this time.

I hope this puts things in perspective from where we stand.

Thanks
Jaime

[florian]

Hi Jaime,

thanks for taking the time to write the long post; I guess I still have a couple of questions for you.

Simply put, after the 11.0.3 server upgrade, about 2-3 weeks ago, the whole system became worst for our requirements.
...
Regardless, the net effect was dissatisfaction

I am pretty surprised to hear this. I won't deny that Scalix 11 overall has given us some serious challenges, and more so than we had expected it to be. This is also why we have decided in early January to basically freeze feature development until the issues are fully resolved. Across the line of 11.0.1, 11.0.2, 11.0.3 we basically only did bugfixes and deep investigations into customer problems. All across the board - and that includes 11.0.3 - our internal testing, our own internal use ('dogfooding') and customer feedback through support and other channels has been favourable of the progress. I actually have no data that would support your experience; i actually don't think any other customer has expressed that opinion to us or had similar experience. Therefore, I would really like to understand in detail what your issues were or are. From our bugtracking, we haven't looked at anything that was working in 11.0.2 and 11.0.3 - therefore we have done no work to address it at this time and it's unlikely that any behaviour that is believed to have seen a change between those versions would change again with 11.0.4. Repeating myself, all changes in 11.0.x have been a series of very specific bug and issue fixes, closely monitored and tested from a development perspective and carefully verified in QA testing and field.

1) being able to utilize Scalix without emails and sometimes whole folders disappearing from OL2003, SWA and even IMAP, as it happened after the 11.0.3 upgrade of the server.

There is no known issue at his point that could cause this. Do you have a reproducible scenario? Did this happen with specific user accounts? Were all the users using various clients at times or did it also happen to users who were using Outlook only or SWA only or IMAP user and none other of the clients? For those using Outlook, were they using SmartCache or not? Did the messages suddenly disappear while the system was running and the user was logged in or did that happen 'overnight'?

2) This SmartCache flip/flop mess also needs a solution. We would like to see the connector 100% problem free when SmartCache is OFF

100% error free in connection with Outlook is a high goal! What exact version (build number) were you using? What kind of issues/errors were you experiencing with SmartCache turned off? Are you using any Outlook plugins? Which AntiVirus solution is configured on those desktops? Are those desktops up to date? What IE version is installed?

and to be able to sync the BlackBerries with the USB cable without the choking of calendar/contact entries anymore (it's happening again!).

What Blackberry device type were you using? Is this happening reproducibly? What version of the Blackberry desktop manager software are you using? How many contacts/calendar entries did the users that were trying the sync have? Did you try with a cached or with a non-cached profile?

3) No more profile recreation. We should be able to switch SmartCache ON/OFF on the fly, even after the installation. SmartCache is not imperative for us at the moment, but we would like to take advantage of improved speed. Should we detect a hint of a problem that could be attributed to SmartCache we can not subject the end user to the intrusion and incovenience of having to delete and create a profile again.

What kind of problem are you experiencing that you believe 'can be attributed to SmartCache'? Based on Outlook connection management and MAPI architecture, it will most likely not become possible to turn SmartCache on and off at runtime - the API does not easily allow such rerouting of the MAPI API calls. We will, however, in the next version of Scalix, introduce functionality that makes it easier to recreate caches should there be a corruption of the cache for any reason.

4) I better developed *upgrade script* (NOT INSTALL), that takes into consideration that a server has been in place from before, and "respects" the customization done previously.

We try hard to preserve configuration. What kind of customization are you specifically talking about?

I can think of a number of other items I'd consider enhancement requests or wish list, but they can wait until 11.0.5 or later. I see no point distracting from the above immediate priorities at this time.

I actually think it is useful to lay things out early; in product software development, we have to think ahead quite a bit. While we have shipped this 11.0.4 today, planning for next month is basically finished and everything is moving along - a lot of work that will make it into the next release(s) has already been done as the last week before a release is more dedicated to QA than to any change. So at this point we are looking at what some of our upcoming releases in Q3 this year would include - therefore providing us with planning input makes sense. Obviously, this should be as specific as possible and for tracking purposes would be more useful in bugzilla enhancement request entries than in forum discussions.

I hope this puts things in perspective from where we stand.

Yes - and no - while I do clearly understand that there are various areas of the product that you are not completely satisfied with, I don't know what specific issues you are running into - hence the questions as per above.

Looking forward to your feedback - and others chiming in to the discussion!

Cheers,
Florian.


Thanks
Jaime[/quote]

[jaime.pinto]

Florian,

Thanks for all feedback.

At the moment I'd like to focus on the upgrade of the server with our reseller to version 11.0.4 over the weekend at some point, possibly even tonight. This way I can install the new connector on my home computer and start doing some tests. We need to fix some of the instabilities since 11.0.3-server.

After that I could post some more findings (and I still have my normal day job to deal with)

Thanks for all the help.
Jaime

[daniel.sand]

just for the book. 11.0.4 Installation worked pretty fine.

[jaime.pinto]

Same here.
11.0.4 server is ok.
Still working on the certification of 11.0.4 connector and SmartCache on OL2003.
Jaime