Offline update FAQ?
Can someone point me to a good cookbook for doing offline updates? My fuzzy understanding is that I would build an internet-accessible SL system, then periodically create my own repository and from that cook a DVD and take it to the non-internet-accessible machine and run Yum against it. I need help filling in the steps. Chris
bash bugs - alternative shell
Hi, I'm wondering if the quickest fix for all of the bash bug stuff coming out now is to replace bash with a different shell. For instance, I see that I have /bin/ksh, why not just link that to /bin/bash and ride out the storm? Is that an alternative? Is there any large subsystem that relies on a bash specific feature?
RE: bash bugs - alternative shell
I happen to have a SL 6.5 system here so I set /bin/sh to point to /bin/dash instead of /bin/bash. The system did not come up in a usable condition. So yes, there are bash specifics in the init scripts. -Original Message- From: Konstantin Olchanski [mailto:olcha...@triumf.ca] Sent: Thursday, October 02, 2014 3:03 PM To: Howard, Chris Cc: SCIENTIFIC-LINUX-USERS Subject: Re: bash bugs - alternative shell On Thu, Oct 02, 2014 at 04:33:17PM +, Howard, Chris wrote: I'm wondering if the quickest fix for all of the bash bug stuff coming out now is to replace bash with a different shell. For instance, I see that I have /bin/ksh, why not just link that to /bin/bash and ride out the storm? Is that an alternative? Is there any large subsystem that relies on a bash specific feature? For interactive use, most people switched from /bin/sh to /bin/tcsh back in the mid-1990-ies. (Bash, ksh, zsh came out much later). For scripting, all shells have bizarre syntax, if your script has if statements or loops, you are better off doing it in perl (or in perl's alternative of the day). -- Konstantin Olchanski Data Acquisition Systems: The Bytes Must Flow! Email: olchansk-at-triumf-dot-ca Snail mail: 4004 Wesbrook Mall, TRIUMF, Vancouver, B.C., V6T 2A3, Canada
RE: bash bugs - alternative shell
Found a script on sourceforge.net - checkbashisms. Downloaded that and ran it against /etc/init.d ugly. Oh well, it was a thought. -Original Message- From: Howard, Chris Sent: Thursday, October 02, 2014 3:34 PM To: 'Konstantin Olchanski' Cc: SCIENTIFIC-LINUX-USERS Subject: RE: bash bugs - alternative shell I happen to have a SL 6.5 system here so I set /bin/sh to point to /bin/dash instead of /bin/bash. The system did not come up in a usable condition. So yes, there are bash specifics in the init scripts. -Original Message- From: Konstantin Olchanski [mailto:olcha...@triumf.ca] Sent: Thursday, October 02, 2014 3:03 PM To: Howard, Chris Cc: SCIENTIFIC-LINUX-USERS Subject: Re: bash bugs - alternative shell On Thu, Oct 02, 2014 at 04:33:17PM +, Howard, Chris wrote: I'm wondering if the quickest fix for all of the bash bug stuff coming out now is to replace bash with a different shell. For instance, I see that I have /bin/ksh, why not just link that to /bin/bash and ride out the storm? Is that an alternative? Is there any large subsystem that relies on a bash specific feature? For interactive use, most people switched from /bin/sh to /bin/tcsh back in the mid-1990-ies. (Bash, ksh, zsh came out much later). For scripting, all shells have bizarre syntax, if your script has if statements or loops, you are better off doing it in perl (or in perl's alternative of the day). -- Konstantin Olchanski Data Acquisition Systems: The Bytes Must Flow! Email: olchansk-at-triumf-dot-ca Snail mail: 4004 Wesbrook Mall, TRIUMF, Vancouver, B.C., V6T 2A3, Canada
RE: File size diff on local disk vs NFS share
I have no idea if this applies to your current situation: Back in the misty past of the unix world it was possible to have a file that contained empty spaces. When a person would then copy that file, the resulting file would be smaller. I believe there was a terminology for this kind of hollow file. But I don't remember what it was and google is not helping so far. Such files were usually binary files associated with complex applications things like databases and such. With the more modern file systems, maybe that situation doesn't happen anymore. Hey! I think I found the terminology: Sparce file. Check out the entry on wikipedia. But that may have nothing to do with your situation. -Original Message- From: aurfalien [mailto:aurfal...@gmail.com] Sent: Thursday, May 03, 2012 12:02 PM To: scientific-linux-us...@fnal.gov Subject: File size diff on local disk vs NFS share Hi all, I never really paid attention to this but a file on an NFS mount is showing 64M in size, but when copying the file to a local drive, it shows 2.5MB in size. My NFS server is hardware Raided with a volume stripe size of 128K were the volume size is 20TB, my local disk is about 500GB. Is this due to my stripe size? Nuggets are appreciated. - aurf
RE: Multiple terminal windows won't send jobs to individual cores
I must have missed where you describe exactly what tasks you are submitting. If you submit multiple CPU-intensive tasks then you should see multiple cores go to high percent used. But if your tasks are I/O bound then the CPU % will not hit 100% as the process block for I/O. Showing 100% CPU may or may not be a good thing. If your tasks were inefficiently programmed they might be CPU bound and possibly some optimization thereafter might overcome that dependency and lead to an I/O bound condition. From: Wil Irwin [mailto:wil.ir...@gmail.com] Sent: Thursday, April 05, 2012 1:24 PM To: zxq9 Cc: Scientific Linux Users List Subject: Re: Multiple terminal windows won't send jobs to individual cores Thanks for all of the suggestions. To recap: 1. I'm not sure why my claim that running jobs in several terminal windows will automatically be distributed to the next least used core (and it will run at 100% until the job is done. I have be doing this for years w/o ANY special coding or using any task managers or scheduler. Not only does (or did) work on SL, but I use this same strategy on an an Ubuntu installation (a virtual Ubuntu running on a Windows machine, to boot!). 2. I am more than willing to look into task managers/schedulers, but I shouldn't need to go down that road. 3. The fact that the tar extraction process is so slow as to be effectively useless, suggest something of a larger problem. 4. These problems exist on 3 machines which are IDENTICAL in every respect and the hardware is less than 1 year old. 5. I will investigate possible BIOS issues; I don't thing dirty power is an issue as the machines are on UPS which provides some degree of power conditioning. And, these units have been in place (about 6 months old) and were in place when things were running correctly. 6. The only hopeful aspect in this very confusing, annoying and frustrating situation is that the behavior is identical on all 3 machines. I welcome other suggestions for troubleshooting. I am hoping I can post a solution to this dilemma in the near future-- my work has come to a screeching halt. Regards, Wil On Thu, Apr 5, 2012 at 11:59 AM, zxq9 z...@zxq9.commailto:z...@zxq9.com wrote: On 04/06/2012 03:18 AM, Stephen John Smoogen wrote: On 5 April 2012 10:47, Wil Irwinwil.ir...@gmail.commailto:wil.ir...@gmail.com wrote: Hi- I am totally stumped and at a complete loss on this one. In an 'old school' manner (a.k.a poor man's grid engine), it is a common practice (at least for me) to open multiple terminal windows on a mullti-core machine. Submitting a job in each terminal window will send it to a core which is not being used. On this particular set of machines I have been doing this for about 2 years. To be honest I have no idea why it worked before. Setting a process to a certain core takes definitive coding to say x will have affinity to CPU y or using a program like taskset to set the affinity. I would try the following: 1) man taskset 2) see if taskset works on your system. Then see if it works. If it doesn't then I would assume that the CPU or some other hardware in the box is having issues and not allowing processes on the other cores for some reason. In addition to the excellent advice here, I've seen a very similar problem before with some faulty BIOS code telling the system to kick in to a very conservative processing mode when certain voltage indicators were met. This came out of the blue and was extremely frustrating to troubleshoot because it like what you're seeing but was a hardware issue born of a special combination of dirty power input, a slowly fading PSU and a cranky mainboard. The odds of that ever happening again anywhere are probably pretty low, especially with server type boards, but its worth considering. Good luck finding your solution. That must be annoying. -z
RE: Multiple Routes
That netmask on the 10.1.0.0 are you sure that covers 10.1.16.x ? I'm not sure the leftmost 19 bytes does the job. But you are doing stuff that I've never done. From: Jeremy Wellner [mailto:jwell...@stanwood.wednet.edu] Sent: Monday, January 30, 2012 1:23 PM To: scientific-linux-us...@fnal.gov Subject: Multiple Routes Hi Guys!
RE: Multiple Routes
Ok, I'm convinced. Sorry to put you through that. Beautifully displayed, BTW. From: Jeremy Wellner [mailto:jwell...@stanwood.wednet.edu] Sent: Monday, January 30, 2012 1:44 PM To: scientific-linux-us...@fnal.gov Subject: Re: Multiple Routes Pretty sure it does. We've gotten our Mac 10.6 servers connected with similar settings. Address: 10.1.0.0 1010.0001.000 0. Netmask: 255.255.224.0 = 19..111 0. Wildcard: 0.0.31.255..000 1. = Network: 10.1.0.0/19http://10.1.0.0/19 1010.0001.000 0. (Class A) Broadcast: 10.1.31.255 1010.0001.000 1. HostMin: 10.1.0.1 1010.0001.000 0.0001 HostMax: 10.1.31.254 1010.0001.000 1.1110 Hosts/Net: 8190 (Private Internethttp://www.ietf.org/rfc/rfc1918.txt) On Mon, Jan 30, 2012 at 12:31 PM, Howard, Chris howa...@prpa.orgmailto:howa...@prpa.org wrote: That netmask on the 10.1.0.0 are you sure that covers 10.1.16.x ? I'm not sure the leftmost 19 bytes does the job. But you are doing stuff that I've never done. From: Jeremy Wellner [mailto:jwell...@stanwood.wednet.edumailto:jwell...@stanwood.wednet.edu] Sent: Monday, January 30, 2012 1:23 PM To: scientific-linux-us...@fnal.govmailto:scientific-linux-us...@fnal.gov Subject: Multiple Routes Hi Guys!
question about SSL proxy solutions
This may be a bit off track for the SL-users, but I'm hoping you can steer me a bit. I have a Dell PE-2950 running Oracle application server. I have the need to use SSL between the desktops and the app server. This is in an intra-net, but with some sensitive data. I'm investigating: 1) turning on SSL in the Oracle app server software This may require a bigger box and maybe more $$ to oracle for licensing. 2) some kind of SSL proxy which would listen, translate and pass along to the app server. Multiple ports are involved. ?- it looks like hardware SSL devices may be primarily for big-pipe super-duper installations and mega-$$ ? We are 100Mbs and not that many users. ?- is there a solution using a second SL box sitting in front of the app server? Thanks for any help you might be pleased to extend. Chris Howard Fort Collins, CO
RE: [SCIENTIFIC-LINUX-USERS] question about SSL proxy solutions
Thanks for all the good info. We are not CPU bound, so I am going to investigate stunnel. Maybe we can use it for other protocols as well as the Oracle app server stuff. That would be very useful. Thanks to all, Chris Howard
running from an external USB drive
I have an external USB drive. I have successfully installed 6.1 (2011-07-27) on the external USB drive. And I it doesn't appear to have touched my internal drive (good!) I previously had done this with 5.3 (?) and used a mkinitrd command and associated cookbook found on a web page (http://www.brunel.ac.uk/~eesridr/extlinux.html) I believe that command puts certain things into a ramdisk, maybe for speedup? I'm not sure what it does. When I try that command under 6.1 I get a warning that it won't override without --force. I didn't do it. Do I want to do it? Things seem to work ok, but I haven't done anything substantial with it yet.
RE: SL6.1 - Unable to login
Usage: -bash {start|stop|restart} - It definitely looks like the style of thing found in /etc/init.d scripts It looks to me like the executable for bash is a link into init.d One way around may be to edit /etc/passwd to make the shell /bin/ksh instead of /bin/bash, then you can get in and look around.
RE: need help with gnutls and p11-kit
Because: 1) this system is a rock-solid Oracle production box that I cannot upgrade without a major effort 2) I'm trying to get ftp-SSL to work on a different box (why ftp-SSL? that would take another 10 steps!) 3) a user can run filezilla from their PC and do the transfer successfully 4) the box I'm really wanting to do the ftp-SSL from has curl 5) curl returns an error 6) curl from this linux box returns the same error 7) network expert says to get filezilla on the linux box to find out of that works. 8) filezilla requres prerequisites which are killing me. 9) some days I wish I was a plumber or a welder. -Original Message- From: Nico Kadel-Garcia [mailto:nka...@gmail.com] Sent: Monday, September 26, 2011 8:19 PM To: Howard, Chris Cc: scientific-linux-us...@fnal.gov Subject: Re: need help with gnutls and p11-kit On Mon, Sep 26, 2011 at 8:12 PM, Howard, Chris howa...@prpa.org wrote: I've been given the task of installing filezilla on SL 5.3 *why*. Updating to SL 5.7, if not SL: 6.1, will save you all sorts of pain.
RE: need help with gnutls and p11-kit
I've got filezilla 3.3.2 running on SL5.2 (gnutls 2.8.5, wxwidgets 2.8.10) all with using stock packages. filezilla-3.3.2-1.el5.rf.i386 (pretty sure that rf is for rpmforge) -- Thanks! I'm making better progress with gnutls 2.8.5 Well, I thought I was. Now I am stuck with a sqlite3 problem.
RE: need help with gnutls and p11-kit
I think I have it working. I had to find the source archive for filezilla and use the older version. -Original Message- From: Mark Stodola [mailto:stod...@pelletron.com] Sent: Tuesday, September 27, 2011 7:05 AM To: owner-scientific-linux-us...@listserv.fnal.gov Cc: 'scientific-linux-us...@fnal.gov' Subject: Re: need help with gnutls and p11-kit Howard, Chris wrote: I've been given the task of installing filezilla on SL 5.3 It demands gnutls with a version higher than what comes with SL. gnutls demands p11-kit (whatever that is!) Compiling p11-kit seems to work but then gnutls can't seem to find it. Configure says No package 'p11-kit-1' found And I don't see any hints about how to resolve that issue. Any help would be appreciated. Chris Howard Chris, I've got filezilla 3.3.2 running on SL5.2 (gnutls 2.8.5, wxwidgets 2.8.10) all with using stock packages. filezilla-3.3.2-1.el5.rf.i386 (pretty sure that rf is for rpmforge) -Mark -- Mr. Mark V. Stodola Digital Systems Engineer National Electrostatics Corp. P.O. Box 620310 Middleton, WI 53562-0310 USA Phone: (608) 831-7600 Fax: (608) 831-9591
RE: need help with gnutls and p11-kit
Probably an overreaction on my part. I've never had a problem with minor level updates either. But I sweat through every time. -Original Message- From: Joshua J. Kugler [mailto:jos...@azariah.com] Upgrades between minor releases (5.x - 5.y) have always been painless for me. Just a matter of yum update. Why would it entail major effort? j
need help with gnutls and p11-kit
I've been given the task of installing filezilla on SL 5.3 It demands gnutls with a version higher than what comes with SL. gnutls demands p11-kit (whatever that is!) Compiling p11-kit seems to work but then gnutls can't seem to find it. Configure says No package 'p11-kit-1' found And I don't see any hints about how to resolve that issue. Any help would be appreciated. Chris Howard
RE: Tips on switching to Oracle Unbreakable Linux
--- original message -- I'm confused as to what you're asking. If you'd like to pay someone for supporting your Linux distribution, then maybe SL have a donation system? I honestly don't understand what criteria you're trying to meet, are you just more comfortable with paying for something that getting something for free? Michael. Chris Howard --- Right now I run yum periodically, then sweat bullets if there are any kernel updates, or just don't run yum.I would rather have someone to yell at if an automatic update breaks things. In theory I could put oracle on a spare box and test it right? Wrong, because licensing for another app server box is many thousands of dollars. The cheapest OUL subscription is something like $200/year and I think that will just let me slurp from their update stream which is all I really want anyway. I've been happy with not doing updates very often. But now I have some (microsoft type) people asking me why I don't have a regular pattern of updates to apply. The best answer is, Because my system actually works. But that is inconvenient in the given circumstances.
RE: Memory limits for Scientific Linux kernels
Do these memory limitations also include shared memory segments? For example, we have Oracle database servers with quite a bit of memory tied up in shared memory segments for the databases. 4 Gig won't get me very far with a large database. Chris Howard
SL and Oracle
I've been an Oracle DBA for quite a number of years, working on HP-UX systems. Now I'm also running some Oracle Application Server on Linux machines and I could use some advice about Oracle and Linux. For SL, I'm assuming I can call my SL 5.3 installation equivalent to Red Hat 5.3 for Oracle support purposes? In the Linux world things seem to move so fast. We have gone years between OS upgrades with HP-UX, and even then there's not a lot of difference as far as Oracle is concerned. I'm checking the Metalink site for a support matrix for various pieces of Oracle Application Server and SL(Redhat) 5.3. Things are all working in this configuration, but I'm nervous that maybe I'm outside the supported setup. Is it safe to do regular yum updates on a dedicated Oracle App Server/SL 5.3 box? Maybe I should just let it run at the current configuration and not chance breaking something. As you can tell, I'm a bit apprehensive about the speed of changes in the Linux world. If anyone knows a good resource for Oracle on Linux, I would be willing to hook up with a consultant for a few moments of good advice.
RE: 5.3 i386 disc7.iso
You are right. Sorry, I should have said v 5.4 That disc #7 is over 700 MB. Chris -Original Message- From: Urs Beyerle [mailto:urs.beye...@env.ethz.ch] Sent: Fri 11/20/2009 1:22 AM To: Howard, Chris Cc: SCIENTIFIC-LINUX-USERS@listserv.fnal.gov Subject: Re: 5.3 i386 disc7.iso Howard, Chris wrote: When I try to burn this one, my burner says it is a DVD image. I think that is because it is too big to fit. My blank CD says it will hold 707 MB. If I buy a different kind of blank CD's will that possibly work? Hi Chris, That's strange. If I download SL.53.031909.CD.i386.disc7.iso from ftp://ftp.scientificlinux.org/linux/scientific/53/iso/i386/CD/ I get an iso file which is 697MB. This should fit on a 700MB CD. Have you checked the md5sum of your downloaded iso file? # wget ftp://ftp.scientificlinux.org/linux/scientific/53/iso/i386/CD/SL.53.031909.CD.i386.disc7.iso # du -csh SL.53.031909.CD.i386.disc7.iso 697MSL.53.031909.CD.i386.disc7.iso 697Mtotal # md5sum SL.53.031909.CD.i386.disc7.iso e6808afa7cd3bdf048b6a0bcdda25088 SL.53.031909.CD.i386.disc7.iso Cheers, Urs
5.3 i386 disc7.iso
When I try to burn this one, my burner says it is a DVD image. I think that is because it is too big to fit. My blank CD says it will hold 707 MB. If I buy a different kind of blank CD's will that possibly work? Chris