From: Bill M. <bm...@ud...> - 2008-02-28 22:15:58
|
I think there may be a bug in Virtualmin 3.53 (GPL). We have 3.46 (GPL) running on one box and it correctly reports domain user counts on virtual-server/index.cgi. The 3.53 installation reports much lower numbers. Checking out the code, it looks like the new version calls &count_domain_users() to get the counts whereas the old version uses the length of &list_domain_users(domain). I'm not 100% sure what causes the erroneous results, but it looks like &count_domain_users expects the home directories to be in a predetermined structure which ours I think don't match. Any thoughts? For now I've changed it use the old method for counting users. |
From: Jamie C. <jca...@we...> - 2008-02-28 22:54:11
|
On 28/Feb/2008 14:15 Bill Moyers wrote .. > I think there may be a bug in Virtualmin 3.53 (GPL). We have 3.46 (GPL) > running on one box and it correctly reports domain user counts on > virtual-server/index.cgi. The 3.53 installation reports much lower > numbers. Checking out the code, it looks like the new version calls > &count_domain_users() to get the counts whereas the old version uses the > length of &list_domain_users(domain). I'm not 100% sure what causes the > erroneous results, but it looks like &count_domain_users expects the home > directories to be in a predetermined structure which ours I think don't > match. Any thoughts? For now I've changed it use the old method for > counting users. Hi Bill, The 3.53 version of Virtualmin changed the way users were counted, as the old method was rather slow - it involved a call to list_domain_users for each domain, which collected a lot more information than was needed. But it is a bug if they don't return the same counts. What kind of directory structure are you using for your users there? - Jamie |
From: John G. <gr...@ag...> - 2008-02-29 00:33:49
|
Jamie Cameron wrote: > On 28/Feb/2008 14:15 Bill Moyers wrote .. > >> I think there may be a bug in Virtualmin 3.53 (GPL). We have 3.46 (GPL) >> running on one box and it correctly reports domain user counts on >> virtual-server/index.cgi. The 3.53 installation reports much lower >> numbers. Checking out the code, it looks like the new version calls >> &count_domain_users() to get the counts whereas the old version uses the >> length of &list_domain_users(domain). I'm not 100% sure what causes the >> erroneous results, but it looks like &count_domain_users expects the home >> directories to be in a predetermined structure which ours I think don't >> match. Any thoughts? For now I've changed it use the old method for >> counting users. >> > > Hi Bill, > > The 3.53 version of Virtualmin changed the way users were counted, as > the old method was rather slow - it involved a call to list_domain_users > for each domain, which collected a lot more information than was needed. > > But it is a bug if they don't return the same counts. What kind of directory > structure are you using for your users there? > > Before we started using virtualmin, the structure for the home directories was /home/<username> It looks like its not counting those users. John |
From: Jamie C. <jca...@we...> - 2008-02-29 05:51:18
|
On 28/Feb/2008 16:33 John Gray wrote .. > Jamie Cameron wrote: > > On 28/Feb/2008 14:15 Bill Moyers wrote .. > > > >> I think there may be a bug in Virtualmin 3.53 (GPL). We have 3.46 (GPL) > >> running on one box and it correctly reports domain user counts on > >> virtual-server/index.cgi. The 3.53 installation reports much lower > >> numbers. Checking out the code, it looks like the new version calls > >> &count_domain_users() to get the counts whereas the old version uses the > >> length of &list_domain_users(domain). I'm not 100% sure what causes the > >> erroneous results, but it looks like &count_domain_users expects the home > >> directories to be in a predetermined structure which ours I think don't > >> match. Any thoughts? For now I've changed it use the old method for > >> counting users. > >> > > > > Hi Bill, > > > > The 3.53 version of Virtualmin changed the way users were counted, as > > the old method was rather slow - it involved a call to list_domain_users > > for each domain, which collected a lot more information than was needed. > > > > But it is a bug if they don't return the same counts. What kind of directory > > structure are you using for your users there? > > > > > Before we started using virtualmin, the structure for the home > directories was /home/<username> > > It looks like its not counting those users. Ok, that would explain it - if the mailbox's home directory is not under the domain's home, it will not be counted. I will fix this in the next Virtualmin release, by falling back to using the GID for determining the domain. - Jamie |
From: John G. <gr...@ag...> - 2008-02-29 13:28:38
|
Jamie Cameron wrote: > Ok, that would explain it - if the mailbox's home directory is not under the > domain's home, it will not be counted. I will fix this in the next Virtualmin > release, by falling back to using the GID for determining the domain. > > - Jamie > > Cool. And I should be sending you patches for the exim support, as well as few other small things in the next day or two. The one that comes to mind is a patch to allow ldap passwords to be encrypted via SSHA. The exim support is coming along. Its detecting exim, can tell if its running, start and stop it, and list the aliases. Adding, modifying, deleting the aliases and domains is mostly coded, but not functional (but its really not very complicated, can't be hard to finish it from here). I wrote and exim module for routines for the exim support in. But its just a lib and some config, no UI. John -- John Gray gr...@ag... AgoraNet, Inc. (302) 224-2475 314 E. Main Street, Suite 1 (302) 224-2552 (fax) Newark, De 19711 http://www.agora-net.com |
From: Jamie C. <jca...@we...> - 2008-02-29 17:12:02
|
On 29/Feb/2008 05:28 John Gray wrote .. > Jamie Cameron wrote: > > Ok, that would explain it - if the mailbox's home directory is not under the > > domain's home, it will not be counted. I will fix this in the next Virtualmin > > release, by falling back to using the GID for determining the domain. > > > > - Jamie > > > > > Cool. And I should be sending you patches for the exim support, as well > as few other small things in the next day or two. The one that comes to > mind is a patch to allow ldap passwords to be encrypted via SSHA. The > exim support is coming along. Its detecting exim, can tell if its > running, start and stop it, and list the aliases. Adding, modifying, > deleting the aliases and domains is mostly coded, but not functional > (but its really not very complicated, can't be hard to finish it from > here). I wrote and exim module for routines for the exim support in. > But its just a lib and some config, no UI. Awesome - SHA support for LDAP passwords will be useful, as a few people have asked about that in the past. - Jamie |