On 2007-05-28 18:23, Carlos E. R. wrote:

The Sunday 2007-05-27 at 11:32 +0200, Richard Bos wrote:

...

...

As I noticed my free space after one year using SUSE is really small i noticed that directories whitch was almost empty after instalation have some size is tehere an easy way to clean it

/var 543MB /tmp 634MB

or if there is some other places where I can find useless files to delete yast -> /etc/sysconfig editor -> system -> cron -> *dirs to clear etc

But don't add /var to that!

Indeed not! /tmp and /var/tmp are the only directories I ever consider need to be cleaned out, but that still leaves a lot of stuff in /var/adm/. I would feel very uncomfortable using a cron task to clean out very much in there, except for /var/adm/backup/rpmdb, which is handled in cron.daily (file /etc/sysconfig/backup). -- Moral indignation is jealousy with a halo. -- HG Wells -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

After I follow the steps on previous email. It seems work on my case. Now my /temp folder are getting lesser in size. Thanks a lot. Tommy On Wed, 2007-05-30 at 10:00 +0800, tino perez wrote:

Too late, now the machine wont boot! :)

Seriously.. what directories are safe to clean?

i too had a problem few-moths ago in where my system run out of space in a few short days after I inadvisable installed a intrusion detection system that created huge log files...

On Tue, 2007-05-29 at 02:23 +0200, Carlos E. R. wrote:

...

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1

The Sunday 2007-05-27 at 11:32 +0200, Richard Bos wrote:

...

...

As I noticed my free space after one year using SUSE is really small i noticed that directories whitch was almost empty after instalation have some size is tehere an easy way to clean it

/var 543MB /tmp 634MB

or if there is some other places where I can find useless files to delete

yast -> /etc/sysconfig editor -> system -> cron -> *dirs to clear etc

But don't add /var to that!

- -- Cheers, Carlos E. R.

-----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.5 (GNU/Linux) Comment: Made with pgp4pine 1.76

iD8DBQFGW3KFtTMYHG2NR9URApsZAJwNa68mBtA1a8Vna12nlKyqEzZ2UgCcCWo1 WRpwn3kWxxl8JSDvuLcgxOA= =ucpy -----END PGP SIGNATURE-----

-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Carlos E. R. wrote:

The Wednesday 2007-05-30 at 10:00 +0800, tino perez wrote:

...

Too late, now the machine wont boot! :)

X-)

...

Seriously.. what directories are safe to clean?

...

i too had a problem few-moths ago in where my system run out of space in

MAX_DAYS_IN_LONG_TMP LONG_TMP_DIRS_TO_CLEAR

I think the LONG refers to long term. These two settings wrong footed me after I upgraded to 10.2. I was a bit puzzled as to why my Windows Profile completely disappeared, (I find using a roaming profile is a useful way of ensuring that some windows stuff gets backed up). I found that the samba profile directories were in the latter, while checking something else out. Now how they got there is a complete mystery grrrr....

-----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.5 (GNU/Linux) Comment: Using GnuPG with SUSE - http://enigmail.mozdev.org iD8DBQFGXVRaasN0sSnLmgIRAuUmAJ97qPOAlJBpfrrz5sm6CRG+rYjbMgCgzJc5 z9jgWhmSU0DXI1XVbfnK6Zc= =YSyE -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Wednesday 2007-05-30 at 19:58 +0200, Richard Bos wrote:

...

  The name "LONG_TMP" of the variables must be indicative of something :-?

LONG as in time. This one can be used for e.g. /var/tmp as that can be populated by users and they may want to have access to their files after some time.

The other varsTMP_DIRS_TO_CLEAR and MAX_DAYS_IN_TMP are meant for e.g. /tmp To remove garbage left behind by system processes.

Still... My /var/tpm contains 17 directories named like "zypp.CU5YqK", containing about 14 KB each, dating since March (system was updated to 10.2 in mid Feb). They seem to contain gpg data. That data should not be there still. There is also a checkinstall.C18067 dir containing 73MB of files of a would be rpm, failed. Neither of those two things are long term needed, I think. Not to mention two files dated 2005 - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.5 (GNU/Linux) Comment: Made with pgp4pine 1.76 iD8DBQFGXgn9tTMYHG2NR9URAizrAJ9oZs2MG/qYIf3/r23nH8Bwcv+DNQCaAlcx a3kJaFstda9K0YKGZH7itag= =gxGR -----END PGP SIGNATURE-----

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Thursday 2007-05-31 at 15:30 +0800, tino perez wrote:

same same here../tmp is full of what looks like yunk and what-not and so-forth.

No, I was commenting about /var/tmp.

how can i be sure that those files are not being use?

lsof. P.S.: Please, do not top post, and trim unneeded text. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.5 (GNU/Linux) Comment: Made with pgp4pine 1.76 iD8DBQFGXpistTMYHG2NR9URAlnMAJ4wIJx+HtbgBvcHz3TPothebBkrTACfX1FQ +ubKRYA3FhCgk0jOxvdJvbw= =K21K -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

I don't know why at first opportunity everyone who replies to any issues, firstly jumps to a command line to accomplish anything. This is not helping new users and works in reverse by concluding that every alteration or fix MUST require a command line fix. We all need to be vary wary of how we are viewed by looking at the following article. It is not appropriate for me to discuss the right or the wrongs of this article - however over 600 other people have expressed their comments and like it or lump it if we do not address the bad things we will never progress form Linux being a Home PC toy to a useful and manageable part of ever day corporate life and if we cannot make that transition we will never be a credible force to be taken seriously. Many of the fixes I see offered to users can be accomplished by the GUI interface. If the interface cannot accomplish the task we need - Then we very seriously need to evaluate where the interface is lacking and we therefore need to log an application Enhancement Bug. For the future we cannot remain with both feet in the ground. If we are totally dependant on a command line entry to correct or set a variable - we have lost already. Remember, MS-Windows did succeed by providing a graphical user interface in total for every beginner to advanced user, (I will exclude admin and config additions and deletions made by admins - NOT users); to be come the worlds largest AND MOST accepted interface that killed DOS, and what finally killed DOS was Windows 95; where at last the user no longer had to contend with editing just 2 files (autoexec.bat, config.sys) Windows 95 was the first true 32bit We cannot deny MS success and willingness to adopt by the world, to do so would argue that Bill Gates is a poor man and MS Windows is not the most prolific interface on available in the world, thus we cannot deny the need to a GUI to perform just about everything for us. Deriving pleasure by writing scripts, or even creating an .RPM file for yourself IS very satisfying for such a user - However is comes at the expense of corporate acceptance and without that Linux will NOT survive or ever threaten Microsoft. This is a horrific crime that such and advanced O/S such as Linux, were we have real memory management issues, where we have a true multitasking O/S, innate security against so many things and where our current GUI is faster and more efficient anything MS could dream of given the same resources. Article - Five crucial things the Linux community doesn't understand about the average computer user http://blogs.zdnet.com/hardware/?p=420&tag=nl.e590 and over 1100 comments can be viewed at http://talkback.zdnet.com/5208-12554-0.html?forumID=1&threadID=34034&messageID=636658 Prophetically speaking, in my lifetime, I hope to see the end of X86 Processor chips. With Intel's multi core (origami) based processor and AMD's 64bit Processors I can see logically medium term winner being the 64 bit AMD type chip. I say this as MS O/S file and memory management is so appalling that major further advancements will require dumping of the architecture. http://www.extremetech.com/article2/0,1697,2000441,00.asp Just a small bit to try and explain my reasons. Please disassociate your concept of the word Domain from the DNS system). Please also acknowledge that we still base our ability to address RAM via the concepts initially created with with first 80386 Intel Processors. This ring principal in inherent to the design of the Microsoft O/S, whether it be DOS or Windows Vista - The latter O/S is still basically running a mutilated form of DOS and Still only able to address very limited amounts of RAM directly and requires an XMS memory management specification to do the juggling up to 4GB (32-bit Version) Windows 95 was the first true 32bit O/S, however no big change was required in processor development as this was accomplished in the inherent design of the first 80386 where the then 16 bit O/S just used 2 clock cycles to address the processor. It was very easy to produce Windows 95 O/S. http://www.extremetech.com/article2/0,1697,2000446,00.asp Further more it is not the architecture that will need changing the 64 bit and 64 bit O/S will no longer function in protected mode but rather Long Mode http://en.wikibooks.org/wiki/X86_Assembly/Protected_Mode This should be a relatively easy task for Linux, but an absolute nightmare for MS. The Long term solution lies is the re-release of the spark and alpha chips and when this happens Linux will just lap it up at the Desktop not only the server. http://www.siliconchip.com.au/cms/A_108523/article.html http://news.zdnet.com/2100-9595_22-826572.html Fundamental security and stability issues of Protected Mode Architecture and O/S utilisation. It is often said that computer systems are insecure because the market demands feature rich, user-friendly software. According to the commonly accepted wisdom, features and convenience are the enemy of security. If it is unrealistic to expect that software complexity will be reduced just to improve security, how can we get out of the break and patch and break cycle? A good place to start would be a redesign of Mainstream Operating Systems (MOSs), both open and closed source, to address some deeply rooted flaws in their protection architecture. Unfortunately, this is unlikely to happen any time soon, but that does not change the fact that it is necessary. The problem with MOSs is their lack of reliable domain separation. You cannot build a secure computing system without reliable domain separation. Domain separation isolates components of the operating system that must work correctly to enforce security, from everything else. The computing pioneers who designed the influential Multics operating system over 30 years ago understood this very well. Multics security was based on a hierarchy of hardware enforced execution domains known as rings. Only a stripped down operating system kernel operated in the most privileged ring (ring 0) while device drivers and other less trusted and potentially buggy operating system components were walled off in rings of lesser privilege. Unlike our current MOSs, in Multics a buggy or malicious driver couldn’t crash the kernel or bypass access controls to steal private data because the processor hardware would stop it from making unauthorised memory accesses. Unfortunately, (at least from the perspective of security) in a MOS there are only two rings. The kernel runs in ring 0 but it shares this domain with device drivers and a huge array of operating system and application components. If you required Administrator rights to install a new application, chances are you added more code that can do whatever it likes because it runs with ring 0 privilege. There is no reliable ‘wall’ between the security critical parts and less trusted parts. This is why an attacker can write code that takes complete control of your computer by exploiting an obscure bug in a sound card driver. Drawing heavily on the Multics approach, Intel x86 processors have supported a four-ring protection architecture since the 286 chip. The problem is that MOSs only use the highest and lowest privilege levels, ignoring the ones in between. This design approach delivers improved performance and makes system and application development easier, but these advantages have been bought at a high price in terms of reduced security. How can the domain separation problem be addressed? The inertia of a huge installed base of insecure systems and the pragmatic need for backward compatibility precludes any solution that involves moving less trusted components out of ring 0. This explains the new processor instructions that Intel announced it would implement to support Microsoft’s NGSCB trusted computing initiative. These new instructions effectively provide a higher privileged ‘ring –1’ to house a hardware protected domain separation mechanism. Intel’s announcement of new processors that support hardware based ‘virtualisation technology’ also shows some promise. Virtualisation technology will allow a MOS to concurrently and transparently share a single processor with another operating system – hopefully, one that implements sound principles of operating system security, principles that have been well understood for over 30 years. *Fundamental to the success of 64-bit Vista will be its ability to not operate in Protected Mode, however I am inclined to think and unable to quantify if is still runs in protected Mode in its 64-bit for, but I would think this highly likely. Longhorne's Server, however I DO believe will run 64-bit and both O/S and Applications will be written to address Long Mode programming in its message addressing and be the most secure and biggest hint at the end of the 80x86 Processor, architecture and the end of the Bastardised DOS form Vists's 32 and 64* bit O/S much to the delight of Linux, Unix and we can just pray all this happens before 2038 ...LOL Good Evening to ALL Scott ;-) 21:41 - GMT +10

I don't know why at first opportunity everyone who replies to  any issues, firstly jumps to a command line to accomplish anything. Not only are you breaking the OT rules here, but you [gasp] hi-jacked a

On Wednesday 13 June 2007 06:42, Registration Account wrote: thread to do it... and I'm not sure how you did that either... since this OT thread was already running elsewhere... :-} -- Kind regards, M Harris <>< -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org