cPanel 11.25 feature discussion

[sehh]

11.25 should then have:

1) roundcube 0.3
2) fixed cpu load when using roundcube
3) actually count the size of the SQL database and add it to the total account usage
4) newer spamassassin
5) ftp quotas that actually work
6) SSL certificates installed in services other than just HTTP (for example, in Exim and Dovecot)

I'm sure it will be out by Christmas time.. it would be a nice present. Those of us in STABLE will probably get it around Easter 2010.

[cpanelkenneth]

This belongs in its own thread.

11.25 should then have:

1) roundcube 0.3

yes

2) fixed cpu load when using roundcube

If using the user-owned roundcube process feature.

3) actually count the size of the SQL database and add it to the total account usage

This is deferred at the moment.

4) newer spamassassin

Yes.

5) ftp quotas that actually work

I'm not familiar with this. Could you expound please?

6) SSL certificates installed in services other than just HTTP (for example, in Exim and Dovecot)

No.

I'm sure it will be out by Christmas time.. it would be a nice present. Those of us in STABLE will probably get it around Easter 2010.

There is far more to cpanel 11.25 than the above. For example look at the Enhancement list in the last few EDGE change logs. The next EDGE build will include even more.

[sehh]

cpanelkenneth, thank you for taking the time to clarify what the new version will include.

If using the user-owned roundcube process feature.

I believe we need a little more information on this issue, I am not aware of such a feature, maybe some documentation will help us enable this when the time comes and 11.25 is released.

This is deferred at the moment.

The database size is VERY important and without this fix many administrators will be disappointed. We definitely need a way to properly calculate the database size, many clients run web applications that store several gigs of data, some webmasters have figured out this vulnerability and use the database as file storage. I can't stress enough how important this fix is and having it deferred is a serious issue.

I'm not familiar with this. Could you expound please?

FTP quotas aren't enforced properly, they are ignored, thus users are able to upload files and overload the storage limit of the hosted account. I opened a bug report many months ago, but I stopped worrying, cpanel has more serious vulnerabilities, like the one above.


About the EDGE build, the changelog is either incomplete or a "new feature" has a completely different meaning for cpanel developers from the rest of us. Looking at the changelog, I don't see anything interesting or amazing that requires a whole year of development. Most changes are either cosmetic, bug fixes or API upgrades. The rest are just version updates from upstream tools.

[VeZoZ]

There is far more to cpanel 11.25 than the above. For example look at the Enhancement list in the last few EDGE change logs. The next EDGE build will include even more.

Where is this? I'm looking at the change logs and the last one was from 2009-08-26 and it's about new upstream versions of software. Before then it was 2009-07-06. Unless this is what you're talking about.

[cpanelkenneth]

1. The RoundCube per-process feature. I've commented on this in other threads. It's an undocumented feature in the current EDGE builds. It's fully documented in the Release Notes for cPanel 11.25.

2. MySQL and PostgreSQL DB Size A new system was put in place to capture the disk usage. The product is only using the DB quantity portion of it at the moment. The new system ( I believe ) is meant to solve two issues:

a. we don't distinguish between database types. For systems that have both PostgreSQL and MySQL, the Max SQL quota in the package is applied in the aggregate. According to feedback, people want it separate ( a different quota for MySQL than for PostgreSQL ). The new system should simplify that.

b. Disk usage. The functionality exists, but is disabled at the moment and is not tied into the reporting system.

3. FTP Quotas I'm not familiar with what you describe. We'll test what you describe against both Pure-FTPd and ProFTPd. Any extra info you could provide would be wonderful. I searched our bugzilla but nothing grabbed my attention.

4. Change Log You are right, there are some entries missing. It's probably due to how some of the features were merged into the product. As for categorization, if fixing something required a redesign to the underlying code, or replaced code entirely ( I don't mean replaced 10 faulty lines with 5 good lines ) then it is categorized as a feature.

Here are some items in the current EDGE builds that I didn't see in the change log:

• Bandwidth tracking and reporting system overhauled.
• Password strength rating system replaced with better algorithm
• ( with the above ) Javascript password rating system uses the same algorithm as the server-side. This means the ratings are consistent
• Non-responsive DNS peers are auto-disabled by the local server
• BIND Restarts trigger by DNS Clustering syncs are deferred to end of syncing
• Task Management system to handle multiple requests for the same event as a single request. For example on a freshly installed server, accessing the Tweak Settings interface and clicking saving will trigger multiple Apache restart requests. These requests are now handed to the task management system (queueprocd) which treats the multiple requests as a single request ( i.e. only one Apache restart is actually issued ).
• Web Template Editor. Allows resellers ( including root ) to modify the web pages displayed for: suspended accounts, the default web page, the account moved page and the connection detection page.

In addition the next EDGE build will have a number of features to prevent XSRF attacks ( some info here: cPanel Security Update: CSRF (cross-site request forgery) - cPanel Inc. ), an improved language system and more.

The Change log entries for RPMs are not what I meant by the "last few EDGE change logs." My apologies. I meant the change logs for the actual builds.

[VeZoZ]

...Task Management system to handle multiple requests for the same event as a single request. For example on a freshly installed server, accessing the Tweak Settings interface and clicking saving will trigger multiple Apache restart requests. These requests are now handed to the task management system (queueprocd) which treats the multiple requests as a single request ( i.e. only one Apache restart is actually issued ).

This sounds interesting but how much does it cover? For example account restores and backups. I believe as it stands now rather than doing a single restart it does one for every subdomain / add-on domain added. So that's a place it make sense to do a single restart after all of them are completed.

There are I'm sure other cases this would be useful as well. So is it used elsewhere as well?

Or better yet is it automated / smart. So all restarts are sent to it and every x interval it groups them together and executes it just once rather than individually as it is now? So it can handle any situation that comes up that cause rolling restarts of Apache.

[Webou]

cpanelkenneth, thank you for taking the time to clarify what the new version will include.




I believe we need a little more information on this issue, I am not aware of such a feature, maybe some documentation will help us enable this when the time comes and 11.25 is released.




The database size is VERY important and without this fix many administrators will be disappointed. We definitely need a way to properly calculate the database size, many clients run web applications that store several gigs of data, some webmasters have figured out this vulnerability and use the database as file storage. I can't stress enough how important this fix is and having it deferred is a serious issue.

I couldn't agree more. This is a must have.

[cpanelkenneth]

This sounds interesting but how much does it cover? For example account restores and backups. I believe as it stands now rather than doing a single restart it does one for every subdomain / add-on domain added. So that's a place it make sense to do a single restart after all of them are completed.

There are I'm sure other cases this would be useful as well. So is it used elsewhere as well?

Or better yet is it automated / smart. So all restarts are sent to it and every x interval it groups them together and executes it just once rather than individually as it is now? So it can handle any situation that comes up that cause rolling restarts of Apache.

At this point only the Apache restarts triggered by saving Tweak Settings are handled by the task manager. It will gradually be extended to handle other Apache restarts, and then to handle other services.

The system is also designed to be used by 3rd party developers. The Task Manager modules come with complete Perl documentation.

More info will be in the release notes. Once those are posted, I will create a link here.

[cPanelDavidG]

Release notes are now available for version 11.25 at: ReleaseNotes < AllDocumentation < TWiki

[cpanelkenneth]

The database size is VERY important and without this fix many administrators will be disappointed. We definitely need a way to properly calculate the database size, many clients run web applications that store several gigs of data, some webmasters have figured out this vulnerability and use the database as file storage. I can't stress enough how important this fix is and having it deferred is a serious issue.

I couldn't agree more. This is a must have.

Please remember I said deferred not dropped. This feature will make an appearance in an 11.25.x version.

[VeZoZ]

Please remember I said deferred not dropped. This feature will make an appearance in an 11.25.x version.

So it's close then? So we may see database size 3 months from now in the edge builds?

[cpanelkenneth]

So it's close then? So we may see database size 3 months from now in the edge builds?

This feature is in EDGE as of build 39565

[DomineauX]

I agree that database disk usage is very important.
I would be ok with databases bypassing hard quota limits as long as it was represented as used space for accounts.
Then we could at least suspend a site that was over quota or contact them about it.

Personally I haven't a clue why separate quotas would be needed for MySQL and Postgres but I could see each account having a normal quota and a database quota.

[robbie_troster]

This feature is in EDGE as of build 39565

I'm not very familiar with the EDGE builds.. I usually run under the Current / Stable builds.

If I change the update to EDGE, will the reliability of the server decrease? are there any bugs that would mess up / create a fault with a web hosting company with a huge client base of around 600+?

[DomineauX]

I'm not very familiar with the EDGE builds.. I usually run under the Current / Stable builds.

If I change the update to EDGE, will the reliability of the server decrease? are there any bugs that would mess up / create a fault with a web hosting company with a huge client base of around 600+?

Yeah you don't want to move to edge on production servers at all.
He was basically letting you know that it is already rolled out in edge so it shall soon make it to current and then release builds.