1. Having the define in the exec allows the Linux admin staff to update what is needed for swap. If an adjustment is made it can be done via the linux level and not have to contact the vm staff to update the profile. Current Linux admin staff do not want to even hear about cms or vm so we try as much as possible to keep them in an area that like.
2. Upping the vdisk space is not the number one option just one of them we would first try to find why then address that.. 3. I'll look into this but the address is always 9999 and the disk ends up /dev/dasd* ends up as the last one all the time even when you add more Disks. Thanks for all the input from every one. This was a first attempt to make some dynamic access and again not have to follow long request paths to get work done by multiple groups (an internal thing) ... Paul -----Original Message----- From: Linux on 390 Port [mailto:[EMAIL PROTECTED] On Behalf Of Mark Post Sent: Thursday, March 06, 2008 6:01 PM To: LINUX-390@vm.marist.edu Subject: Re: Swap space problems... >>> On Thu, Mar 6, 2008 at 2:50 PM, in message <[EMAIL PROTECTED]>, "Ayer, Paul W" <[EMAIL PROTECTED]> wrote: > In an effort to not have to have the Linux Admin staff do something in > VM I added the exec below to the boot process. > I'm still fine tuning it but over all it works rather well. Why would your Linux sysadmins be doing "something in VM" when you can just do it for them in a standard PROFILE EXEC and not have to modify your Linux systems? -snip- > Monitoring should then be done to the real disk swap space to see if > it's being used. > If it is we were told that we would then want to make the vdisk space > larger and also tune the system as well. Making the VDISK bigger is probably not the correct change to make. You may want to add a second VDISK or modify the guest's virtual storage instead. Making the VDISK bigger just puts more pressure on XSTOR because of the way Linux uses its paging spaces. > Anyway here is the exec that is run at boot time; -snip- > echo 1 > /sys/bus/ccw/drivers/dasd-fba/0.0.9999/online If you insist on doing this in Linux, you may want to consider changing this to use the dasd_configure or chccwdev command instead. Less likely to break in the future. You also should turn on DIAG I/O for better performance. > sleep 10 > cat /proc/dasd/devices | grep 9999 | cut -c35-40 > /tmp/device > vswap=$(cat /tmp/device) > mkswap /dev/$vswap > swapon -p50 /dev/$vswap This is trouble waiting to happen. You would be better off making the device number a fixed entry in your DASD parameter string so that it would always have the same device name. Dynamically issuing mkswap commands against a derived device name from a file that wasn't deleted since the last boot is more danger than I would care to court. Mark Post ---------------------------------------------------------------------- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 ---------------------------------------------------------------------- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
