ahsu@grad1.cis.upenn.edu (Andrew C Hsu) (02/04/91)
Hello, I've been getting this problem on my RS6000 where I have to reboot the machine about once a week because the paging space gets low. Currently I've installed 48Megs of paging space over 2 physical volumes. Is there a way to reset the paging space with having to reboot? And is this problem I'm having a bug with AIX3.1? --> Andrew Hsu
javey@hybrid.UUCP (shahram javey) (02/06/91)
In article <36829@netnews.upenn.edu> ahsu@grad1.cis.upenn.edu (Andrew C Hsu) writes: >Hello, > >I've been getting this problem on my RS6000 where I have to reboot the >machine about once a week because the paging space gets low. Currently >I've installed 48Megs of paging space over 2 physical volumes. Is there is this 48Meg per phyical vol.? If you're running X and using xlc you may need 48Meg on each phy. vol. >a way to reset the paging space with having to reboot? And is this login as root and use smit to change the paging space. you should not need to reboot. I've done it without rebooting.
robin@batcomp.austin.ibm.com (Robin D. Wilson) (02/08/91)
In article <36829@netnews.upenn.edu> ahsu@grad1.cis.upenn.edu (Andrew C Hsu) writes: >Hello, > >I've been getting this problem on my RS6000 where I have to reboot the >machine about once a week because the paging space gets low. Currently >I've installed 48Megs of paging space over 2 physical volumes. Is there >a way to reset the paging space with having to reboot? And is this >problem I'm having a bug with AIX3.1? At 3001 update levels and below, there was a problem with paging space on machines that used the hft device (like the large color monitor). This problem is mostly corrected in 3002, but not completely solved until the 3003 update. 3003 should be available soon, so I suggest that you wait it out, and just reboot the machine frequently. -- +-----------------------------------------------------------------------------+ |The views expressed herein, are the sole responsibility of the typist at hand| +-----------------------------------------------------------------------------+ |UUCP: robin%aixserv@uunet.uu.net | |USNail: 701 Canyon Bend Dr. | | Pflugerville, TX 78660 | | Home: (512)251-6889 Work: (512)823-3015 | +-----------------------------------------------------------------------------+
jfh@greenber.austin.ibm.com (John F Haugh II) (02/08/91)
In article <36829@netnews.upenn.edu> ahsu@grad1.cis.upenn.edu (Andrew C Hsu) writes: >I've been getting this problem on my RS6000 where I have to reboot the >machine about once a week because the paging space gets low. Currently >I've installed 48Megs of paging space over 2 physical volumes. Is there >a way to reset the paging space with having to reboot? And is this >problem I'm having a bug with AIX3.1? There were problems in the kernel (notably X-windows & hft's) involving page space losage. It sounds as though you have an old level of AIX v3 and need to upgrade. -- John F. Haugh II | I've Been Moved | MaBellNet: (512) 838-4340 SneakerNet: 809/1D064 | AGAIN ! | VNET: LCCB386 at AUSVMQ BangNet: ..!cs.utexas.edu!ibmchs!auschs!snowball.austin.ibm.com!jfh (e-i-e-i-o)
jfh@rpp386.cactus.org (John F Haugh II) (02/10/91)
In article <5170@awdprime.UUCP> robin@batcomp.austin.ibm.com (Robin D. Wilson) writes: >At 3001 update levels and below, there was a problem with paging space on >machines that used the hft device (like the large color monitor). This >problem is mostly corrected in 3002, but not completely solved until the >3003 update. 3003 should be available soon, so I suggest that you wait it >out, and just reboot the machine frequently. Or you could avoid unnecessarily killing the X server, which causes the loss of quite a few pages at a whack. That was the biggest page sucker that I was ever aware of. -- John F. Haugh II UUCP: ...!cs.utexas.edu!rpp386!jfh Ma Bell: (512) 832-8832 Domain: jfh@rpp386.cactus.org "I've never written a device driver, but I have written a device driver manual" -- Robert Hartman, IDE Corp.
jet@karazm.math.uh.edu ("J. Eric Townsend") (02/10/91)
In article <19044@rpp386.cactus.org> jfh@rpp386.cactus.org (John F Haugh II) writes: >Or you could avoid unnecessarily killing the X server, which causes the >loss of quite a few pages at a whack. That was the biggest page sucker >that I was ever aware of. I have a question about paging/swapping, and I hope that someone here is legally able to answer it. This is from an RS/6000, model 320, 8Mb RAM, 30+MB paging, AIX 3002, just rebooted. $ ps aux USER PID %CPU %MEM SZ RSS TT STAT TIME CMD root 0 0.0% 78% 7536 6360 - S 0:01 swapper <rest of listing deleted for brevity> The manual page for ps says that SZ is... Interesting, I just did "man ps" and got back a prompt. "man -k ps" shows that there is indeed a man page for ps. Oh well. You probably know what SZ and RSS mean anyway... The point is, why is the swapper taking up 78% of memory?!? I called IBM Defective Support and asked if it should be that large, and the convo was (paraphrased): me: "Is it supposed to be that big? If so, what does it mean? Other unixes don't behave this way." support: "yes, it's supposed to be that big." me: "Well, is that the size of the page table perhaps, and ps(1) is confused, and believes that swapper is that big, when in fact, swapper just happens to own *all* the pages for the system?" support: "No, the swapper is just big. AIX does things differently than other unixes" We're going to go to 32Mb soon (finally). Will 6-7Mb of our memory be wasted by the swapper then as well? We have a typical job size of 25-30Mb, so we feel we should only have to go to 32Mb -- the OS shouldn't eat up more than 2Mb of space if there's only one user process running. This is something we are *very* concerned about, as we're considering buying several 8Mb machines and using them as computation engines for small jobs (ie: installing only the os and compiler packages, no windowing software, get there via telnet or dumb terminal). If we have to buy an extra 8Mb of ram just for the OS, then the cost becomes unreasonable. Answers? Hints? Anonymous phone calls with an adb-able patch? :-) -- J. Eric Townsend - jet@uh.edu - bitnet: jet@UHOU - vox: (713) 749-2120 "It is the cunning of form to veil itself continually in the evidence of content. It is the cunning of the code to veil itself and to produce itself in the obviousness of value." -- Baudrillard