Submitted by tbirnseth on Wed, 07/01/2009 - 02:06
I upgraded my test system and all went fine (I think). I then went to upgrade my production server and got the following errors from usermin install. Please advise.
Installing package(s) with command yum -y install usermin .. Loaded plugins: downloadonly, rhnplugin, security Excluding Packages in global exclude list Finished Setting up Install Process Parsing package install arguments Resolving Dependencies --> Running transaction check ---> Package usermin.noarch 0:1.410-1 set to be updated Traceback (most recent call last): File "/usr/bin/yum", line 29, in ? yummain.user_main(sys.argv[1:], exit_code=True) File "/usr/share/yum-cli/yummain.py", line 229, in user_main errcode = main(args) File "/usr/share/yum-cli/yummain.py", line 145, in main (result, resultmsgs) = base.buildTransaction() File "/usr/lib/python2.4/site-packages/yum/__init__.py", line 647, in buildTransaction (rescode, restring) = self.resolveDeps() File "/usr/lib/python2.4/site-packages/yum/depsolve.py", line 704, in resolveDeps for po, dep in self._checkFileRequires(): File "/usr/lib/python2.4/site-packages/yum/depsolve.py", line 939, in _checkFileRequires if not self.tsInfo.getOldProvides(filename) and not self.tsInfo.getNewProvides(filename): File "/usr/lib/python2.4/site-packages/yum/transactioninfo.py", line 414, in getNewProvides for pkg, hits in self.pkgSack.getProvides(name, flag, version).iteritems(): File "/usr/lib/python2.4/site-packages/yum/packageSack.py", line 300, in getProvides return self._computeAggregateDictResult("getProvides", name, flags, version) File "/usr/lib/python2.4/site-packages/yum/packageSack.py", line 470, in _computeAggregateDictResult sackResult = apply(method, args) File "/usr/lib/python2.4/site-packages/yum/sqlitesack.py", line 861, in getProvides return self._search("provides", name, flags, version) File "/usr/lib/python2.4/site-packages/yum/sqlitesack.py", line 43, in newFunc return func(*args, **kwargs) File "/usr/lib/python2.4/site-packages/yum/sqlitesack.py", line 837, in _search for pkg in self.searchFiles(name, strict=True): File "/usr/lib/python2.4/site-packages/yum/sqlitesack.py", line 43, in newFunc return func(*args, **kwargs) File "/usr/lib/python2.4/site-packages/yum/sqlitesack.py", line 568, in searchFiles self._sql_pkgKey2po(rep, cur, pkgs) File "/usr/lib/python2.4/site-packages/yum/sqlitesack.py", line 470, in _sql_pkgKey2po pkg = self._packageByKey(repo, ob['pkgKey']) File "/usr/lib/python2.4/site-packages/yum/sqlitesack.py", line 413, in _packageByKey po = self.pc(repo, cur.fetchone()) File "/usr/lib/python2.4/site-packages/yum/sqlitesack.py", line 68, in __init__ self._read_db_obj(db_obj) File "/usr/lib/python2.4/site-packages/yum/sqlitesack.py", line 94, in _read_db_obj setattr(self, item, _share_data(db_obj[item])) TypeError: unsubscriptable object .. install failed!
Status:
Closed (fixed)
Comments
http://www.virtualmin.com/node/10164#comment-43993
Submitted by tbirnseth on Wed, 07/01/2009 - 11:44 Comment #2
What does the above comment mean? It says it's fixed but no info on what I'm supposed to do or what was found....
Submitted by tbirnseth on Wed, 07/01/2009 - 11:46 Comment #3
Okay, got it. It didn't look like a link!
All this bold text is strange.
Submitted by andreychek on Wed, 07/01/2009 - 13:16 Comment #4
Yeah, there was some trouble with the repository in the past, and the fix is the command that you ran, which clears out all the metadata.
Sounds like it's working now -- that's great!
Submitted by tbirnseth on Wed, 07/01/2009 - 13:23 Comment #5
Yes, went fine.
Unfortunately it installed the new Apache which has broken all the basic_authentication. So my system is now unusable by my clients.
Is there a sane way to revert to the prior version of Apache via Virtualmin (and knowing there won't be any 'gotchas')?
But the new Apache does not seem to populate $_SERVER['PHP_AUTH_USER'] and $_SERVER['PHP_AUTH_PW'] anymore on a secure server when this info is passed in the Authentication_basic header.
Right now all my clients are down... I know it's not a VM issue but thought you might have some insight!
Submitted by andreychek on Wed, 07/01/2009 - 14:49 Comment #6
Well, poking through the PHP manual, it looks like those variables are set when you use mod_php.
If you've switched to using PHP through CGI, that would no longer be populated.
Submitted by tbirnseth on Wed, 07/01/2009 - 15:25 Comment #7
Well, it worked before I upgraded to the new Apache via the htaccess excerpt above.... Now all my clients are failing and I have to completely redo the security in a rush mode.... Not a pleasant situation. If you have any thoughts on why this would have changed, I'm all ears!
All this bold text is strange.
What do you mean? What browser are you using?
Well, it worked before I upgraded to the new Apache via the htaccess excerpt above.... Now all my clients are failing and I have to completely redo the security in a rush mode.... Not a pleasant situation. If you have any thoughts on why this would have changed, I'm all ears!
I'm confused. I don't see anything about htaccess in this ticket. Are you running Usermin under Apache? Or is this a different problem?
Submitted by tbirnseth on Wed, 07/01/2009 - 16:10 Comment #10
Firefox. All the text other than what's quoted is bold (even this as I type).
That's really weird. I'm not seeing that. I mean...the "ISSUE TITLE" and such is all bold. But, the comments are all non-bold and reasonable sized. What specific OS/version are you running on your client machine? I'd obviously like to fix this, as it sounds really strange and annoying. ;-)
Maybe attach a screenshot, so I can see what I'm looking for?
Submitted by tbirnseth on Wed, 07/01/2009 - 16:22 Comment #12
XP SP2
I tried to attach a screen shot in a word document but got a javascript error of: An HTTP error 0 occurred. /comment-upload/js
Heheh...I'm sure Word docs are not in the list of accepted file types for attachments. jpg, png, gif, etc. should work fine.
I have an XP SP2 system in the closet, I'll check the site on that, so see if I can figure out what's happening.
Submitted by tbirnseth on Wed, 07/01/2009 - 16:24 Comment #14
Hmm it shows in my "Comment #5".
The problem is not in VM, or any of your products. The problem is in Apache.
You guys are pretty smart so thought you might have heard complaints from others or know of a solution.
Submitted by tbirnseth on Wed, 07/01/2009 - 16:26 Comment #15
Also, I seem to get 2 emails for each response you send. One has no comments, just the link to the message (it's the 1st one) and the second has the comments you made. Can you get rid of the first one? I'm guessing it's because some status changed or something.....
Very minor issue....
Submitted by andreychek on Wed, 07/01/2009 - 16:40 Comment #16
Alright, so, where are we at -- are you still having Apache/PHP problems?
I'm not sure why it happened, but it's due to mod_php going away.
When is the last time you're certain that it worked?
The key may be to simply re-enable mod_php, but again, I'm not sure why it stopped working.
If you aren't sure how to do that, I can take a look.
Also, I seem to get 2 emails for each response you send.
So, I've seen this, too... but in my case I had double subscriptions (both a type and tag that matched, or in the forums, I was subscribed to both the category and the forum). I don't see why you would be getting double messages though. Your subscriptions look exactly right, and not doubled up.
Status changes would and should generate messages, even if they have no content. But, the status of this ticket has only changed a couple of times during its life, so if you're always getting doubles, something is definitely wrong.
Submitted by andreychek on Wed, 07/01/2009 - 16:43 Comment #18
Oh, and I meant to add -- for future reference, I recommend opening a new ticket for new issues.
This ticket relates to Usermin... so we get thrown off a bit when the topic turns to a mod_php problem :-)
We're happy to help with issues that arise, but opening a new ticket would be the way to go. Thanks!
Submitted by tbirnseth on Wed, 07/01/2009 - 18:25 Comment #19
It stopped working this morning right after I upgraded to Apache. That's when my phone started ringing.
I know it's not your issue, it's mine for being dumb enough to change versions! Not again!
I'm guessing they removed the solution about putting it in the environment to get around the cgi shortcoming on a secure server.
This server has always run mod_fcgi.
I just have no idea how to get the functionality back again! Was hoping you'd have some insight.
tony P.S. I got two messages before this one with only my original message and the link to this thread.
Submitted by andreychek on Wed, 07/01/2009 - 21:41 Comment #20
Okay, is there any chance I could log in and take a look?
Also, can you give me an example of a domain that's not longer working?
Thanks!
Submitted by tbirnseth on Wed, 07/01/2009 - 21:54 Comment #21
Sure: Since I can't see where to make this private, I sent the credentials to joe@virtualmin.com.
It's any domain. The issue is that I use (have tweaked my critical pieces now) Basic Authentication when sending XML data to the server. The user/password is what's used to identify which company within the application the xml data is destined for.
I wrote a little app that will demonstrate the problem. It's in /home/ezom/public_html/om/x.php If it's installed on an older Apache server, tweak the $nonSSL and $SSL URL's to appropriate places. The nonSSL should always fail and the SSL should always succeed.
My test server is non-ssl and my production server is ssl.
I know this is an Apache issue, not VM. I was hoping someone would know what changed so I could figure out how to fix it. Ultimately I will rework all my little applets to include credentials within the document versus relying on Basic Authentication, but I'm not there yet. However, the most critical are addressed.
Appreciate your insight/help in this.
tony
Submitted by andreychek on Wed, 07/01/2009 - 22:00 Comment #22
Thanks -- can you send that info to eric@virtualmin.com as well?
Joe may be able to help, but I don't know that he's sitting around his computer right now ;-)
Submitted by andreychek on Thu, 07/02/2009 - 14:06 Comment #23
We discussed/resolved the remaining Apache issue via email.
Long story short, thank goodness for backups :-)
Submitted by tbirnseth on Thu, 07/02/2009 - 15:47 Comment #24
Just thought I'd add the result.
Problem was that my .htaccess got overwritten somehow and so the RewriteEngine On directive was missing. Restoring my original .htaccess makes it work perfectly.
Submitted by Issues on Thu, 07/16/2009 - 17:18 Comment #25
Automatically closed -- issue fixed for 2 weeks with no activity.