Hello list, There are some days I've configured 'phpMyAdmin' as comented in '.../phpMyAdmin/Documentation.html', and I've try it. All was working fine till the moment I've try to logout (from phpMyAdmin session): no way. From this point I was unable to run apache, not only phpMyAdmin. After trying to access my main page (pure html), in the /var/log/httpd/error.log I've: .... PHP Fatal error: Unable to start session mm module in Unknown on line 0 PHP Fatal error: Unable to start session mm module in Unknown on line 0 ..... I've unloaded mod_php4, then the message was: .... Wed Apr 16 18:06:08 2003] [notice] child pid 8006 exit signal Segmentation fault (11) [Wed Apr 16 18:54:17 2003] [notice] child pid 8011 exit signal Segmentation fault (11) ..... Somebody knows how to unlock that? On the other hand, today I've seen the last SuSE 8.1 update for glibc is 2.2.5-161, but surprisingly YOU tell me my instaled version is 2.2.5-164 (installed on 9.Apr.2003, I think via YOU because I've no .rpm file, nor .gz on my system), and now it seems this update don't exists. Somebody in SuSE knows about this?. I don't know if there are some relation between this two issues, but phpMyAdmin uses 'iconv' for translations of charsets, that is included on glibc package.... Thanks in advance (...I'll cut mine veins with the tape of the printer....)
On Wednesday 16 April 2003 21:26, Pedro Cáliz wrote:
..... On the other hand, today I've seen the last SuSE 8.1 update for glibc is 2.2.5-161, but surprisingly YOU tell me my instaled version is 2.2.5-164 (installed on 9.Apr.2003, I think via YOU because I've no .rpm file, nor .gz on my system), and now it seems this update don't exists. Somebody in SuSE knows about this?. ....
Hello list, I've seen there exist releases 164 and 165 for glibc_2.2.5 in ftp.suse.com, but my YOU reports release 161 as the newest one, and I don't understand why. Must I update this package by hand? Releases 164/165 aren't recommended? Are there more problems with yast/you? (I've updated this packages too, but there are no changes) Regards Pedro
On Apr 22, Pedro Cáliz
I've seen there exist releases 164 and 165 for glibc_2.2.5 in ftp.suse.com, but my YOU reports release 161 as the newest one, and I don't understand why. Because one is i586 and one is i686. Unfortunately the update description file format is not very good to handle different architectures. I've sent a mail to feedback@suse.com with a suggestion for improvement but haven't received an answer yet :( If anybody could give me a direct contact ...
Markus -- __________________ /"\ Markus Gaugusch \ / ASCII Ribbon Campaign markus@gaugusch.at X Against HTML Mail / \ http://fou4s.gaugusch.at/ - Fast OnlineUpdate for SUSE
Hi Markus,
Kind of you this super-fast response: six minutes!!
Have you contacted Thomas Biege
Because one is i586 and one is i686. Unfortunately the update description file format is not very good to handle different architectures. I've sent a mail to feedback@suse.com with a suggestion for improvement but haven't received an answer yet :( If anybody could give me a direct contact ...
Markus
On Wed, Apr 16, 2003 at 09:26:17PM +0200, Pedro Cáliz wrote:
There are some days I've configured 'phpMyAdmin' as comented in '.../phpMyAdmin/Documentation.html', and I've try it. All was working fine till the moment I've try to logout (from phpMyAdmin session): no way. From this point I was unable to run apache, not only phpMyAdmin. After trying to access my main page (pure html), in the /var/log/httpd/error.log I've:
.... PHP Fatal error: Unable to start session mm module in Unknown on line 0 PHP Fatal error: Unable to start session mm module in Unknown on line 0 .....
What does the command 'cat /proc/sys/kernel/shmmax' output?
I've unloaded mod_php4, then the message was:
"unloaded"? What do you mean?
.... Wed Apr 16 18:06:08 2003] [notice] child pid 8006 exit signal Segmentation fault (11) [Wed Apr 16 18:54:17 2003] [notice] child pid 8011 exit signal Segmentation fault (11) .....
Peter -- Thought is limitation. Free your mind.
Hello Peter, Your efforts are really welcome. I'm very worried about this issue. On Thursday 24 April 2003 11:47, Peter Poeml wrote:
.....
What does the command 'cat /proc/sys/kernel/shmmax' output?
'cat /proc/sys/kernel/shmmax' -- > 26458112
I've unloaded mod_php4, then the message was:
"unloaded"? What do you mean?
You know there are some modules added by the suse_loadmodule.conf file, one of wich is mod_php4. Well, I've get off this module from that file (with yast), so mod_php isn't loaded with apache start. This morning I've wipe all the phpMyAdmin, mod_php4, mod_php4-core and phplib packages, and I've reinstalled 'cleanly' mod_php4, mod_php4-core and phplib. But the problem remains. I've no idea where may reside the small change that locks all, and there are no other message than 'PHP Fatal error: Unable to start session mm module in Unknown on line 0' in the /var/log/httpd/error.log. Do you know if apache has a 'debug' or 'trace all' mode? Regards Pedro
On Thu, Apr 24, 2003 at 05:06:39PM +0200, Pedro Cáliz wrote:
On Thursday 24 April 2003 11:47, Peter Poeml wrote:
.....
What does the command 'cat /proc/sys/kernel/shmmax' output?
'cat /proc/sys/kernel/shmmax' -- > 26458112
Please try again to start apache after issueing 'echo 33554432 > /proc/sys/kernel/shmmax'. Do you have powertweak installed, and is it started by default? It is a known bug that it lowers the value in a default configuration.
I've unloaded mod_php4, then the message was:
"unloaded"? What do you mean?
You know there are some modules added by the suse_loadmodule.conf file, one of wich is mod_php4. Well, I've get off this module from that file (with yast), so mod_php isn't loaded with apache start.
For Apache1, you would need to get rid of _two_ commands: the LoadModule command and the AddModule command (which in the other suse_* file). The best way is to let SuSEconfig (SuSEconfig --module apache) generate those files, after modifying the variables in /etc/sysconfig/apache.
This morning I've wipe all the phpMyAdmin, mod_php4, mod_php4-core and phplib packages, and I've reinstalled 'cleanly' mod_php4, mod_php4-core and phplib. But the problem remains. I've no idea where may reside the small change that locks all, and there are no other message than 'PHP Fatal error: Unable to start session mm module in Unknown on line 0' in the /var/log/httpd/error.log. Do you know if apache has a 'debug' or 'trace all' mode?
Yes, set LogLevel in /etc/httpd/httpd.conf to debug, but it won't show more I guess. The problem in your case is that php won't start (not apache) because it wants to allocate a large amount of shared memory by default. Try 'echo | php' to see this confirmed... Peter -- Thought is limitation. Free your mind.
On Friday 25 April 2003 00:32, Peter Poeml wrote:
On Thu, Apr 24, 2003 at 05:06:39PM +0200, Pedro Cáliz wrote:
On Thursday 24 April 2003 11:47, Peter Poeml wrote:
.....
What does the command 'cat /proc/sys/kernel/shmmax' output?
'cat /proc/sys/kernel/shmmax' -- > 26458112
Please try again to start apache after issueing 'echo 33554432 > /proc/sys/kernel/shmmax'.
GREAT!!! This appears to be enough. All is working again.
Do you have powertweak installed, and is it started by default? It is a known bug that it lowers the value in a default configuration.
Yes, I've it installed since 6.Apr.2003, but all was working fine till 12.04.2003. (???)
I've unloaded mod_php4, then the message was:
"unloaded"? What do you mean?
You know there are some modules added by the suse_loadmodule.conf file, one of wich is mod_php4. Well, I've get off this module from that file (with yast), so mod_php isn't loaded with apache start.
For Apache1, you would need to get rid of _two_ commands: the LoadModule command and the AddModule command (which in the other suse_* file). The best way is to let SuSEconfig (SuSEconfig --module apache) generate those files, after modifying the variables in /etc/sysconfig/apache.
I do all changes with yast, I'm not really an expert to doing so by hand (yet...I hope... ;-) )
This morning I've wipe all the phpMyAdmin, mod_php4, mod_php4-core and phplib packages, and I've reinstalled 'cleanly' mod_php4, mod_php4-core and phplib. But the problem remains. I've no idea where may reside the small change that locks all, and there are no other message than 'PHP Fatal error: Unable to start session mm module in Unknown on line 0' in the /var/log/httpd/error.log. Do you know if apache has a 'debug' or 'trace all' mode?
Yes, set LogLevel in /etc/httpd/httpd.conf to debug, but it won't show more I guess.
Interesting for future issues.
The problem in your case is that php won't start (not apache) because it wants to allocate a large amount of shared memory by default. Try 'echo | php' to see this confirmed...
Peter
Thanks a lot. I'm sure I'll never be able to find the solution my shelf. Regards Pedro
Hello Peter, Your recommendation works fine till the systems reboots: /proc/sys/kernel/shmmax has again the value 26458112. I've take a look in powertweak, and I can find two params. about shared memory: - kernel-shmall = 2 - kernel/shmall = 2097152 Both entries sugest that, with some products installed like Oracle, you must raise this values. Do you think I must do it? Regards, Pedro
On Fri, Apr 25, 2003 at 01:16:33PM +0200, Pedro Cáliz wrote:
Your recommendation works fine till the systems reboots: /proc/sys/kernel/shmmax has again the value 26458112.
I've take a look in powertweak, and I can find two params. about shared memory: - kernel-shmall = 2 - kernel/shmall = 2097152
Both entries sugest that, with some products installed like Oracle, you must raise this values. Do you think I must do it?
For some reason the powertweak package ships with a strange default value for this parameter. You've got two choices: deinstall (or not start) powertweak, or play with that setting so a value of ~32 MB is restored. (33554432 is 0x2000000) Peter -- Thought is limitation. Free your mind.
On Tuesday 29 April 2003 10:14, Peter Poeml wrote:
On Fri, Apr 25, 2003 at 01:16:33PM +0200, Pedro Cáliz wrote:
Your recommendation works fine till the systems reboots: /proc/sys/kernel/shmmax has again the value 26458112.
I've take a look in powertweak, and I can find two params. about shared memory: - kernel-shmall = 2 - kernel/shmall = 2097152
Both entries sugest that, with some products installed like Oracle, you must raise this values. Do you think I must do it?
For some reason the powertweak package ships with a strange default value for this parameter. You've got two choices: deinstall (or not start) powertweak, or play with that setting so a value of ~32 MB is restored. (33554432 is 0x2000000)
Peter
Withdrawall is the last ;-) , so I try to correct it. 'kernel-shmall' rounds decimal values (all I need is 2,5), then I've change it to 3 and now /proc/sys/kernel/shmmax has a value of 39687168 ( about 37,8 MB). Thanks a lot, Pedro
participants (3)
-
Markus Gaugusch -
Pedro Cáliz -
Peter Poeml