[opensuse] How to replace drive containing /swap ??
Listmates, Critical situation. Server running 100G drive. The drive started to fail. Plenty of backups available and I was able to recovered all data to another server. Installed fresh 250G sata drives configured in hardware raid (sdb and sdc). Used the /swap from the failing drive (sda) for install purposes because I didn't was to put swap on the raid. Fresh 10.3 install replacing last mandrake server went fine. Server runs fine, just gives occasional /sda offline errors. Problem, the old drive with /swap is sda, the /swap partition is sda5. The sda drive must be replaced, but how? The question: "Can I just install a new sda, and boot and then partition a new /swap with yast, or, will the boot process freak out when /swap isn't available and cause the boot to fail?" I'm going to try it, but if someone else has done this and has advise to help avoid any "gotchas", I would definitely appreciate a heads up. Thanks! -- David C. Rankin, J.D., P.E. Rankin Law Firm, PLLC 510 Ochiltree Street Nacogdoches, Texas 75961 Telephone: (936) 715-9333 Facsimile: (936) 715-9339 www.rankinlawfirm.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
David C. Rankin a écrit :
Listmates,
Critical situation. Server running 100G drive. The drive started to fail. Plenty of backups available and I was able to recovered all data to another server. Installed fresh 250G sata drives configured in hardware raid (sdb and sdc). Used the /swap from the failing drive (sda) for install purposes because I didn't was to put swap on the raid. Fresh 10.3 install replacing last mandrake server went fine. Server runs fine, just gives occasional /sda offline errors.
Problem, the old drive with /swap is sda, the /swap partition is sda5. The sda drive must be replaced, but how?
The question: "Can I just install a new sda, and boot and then partition a new /swap with yast, or, will the boot process freak out when /swap isn't available and cause the boot to fail?" I'm going to try it, but if someone else has done this and has advise to help avoid any "gotchas", I would definitely appreciate a heads up.
Thanks!
if you have enough ram, stop the swap (swapoff), unset the swap file in fstab (write #on beginning of the swap line). no swap is really needed after that set any swap file on any disk (make swap, swap on) and re-enable the swap file jdd -- http://www.dodin.net -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
jdd wrote:
if you have enough ram, stop the swap (swapoff), unset the swap file in fstab (write #on beginning of the swap line).
no swap is really needed
after that set any swap file on any disk (make swap, swap on) and re-enable the swap file
jdd
Thanks! There was 2G of ram and the server is under no load as it is being rebuilt so it never swaps. I had an old (very healthy) 20G drive. I just unplugged the cables from the 100G, plugged them into the 20G (laying in the bottom of the box for testing) and booted. System came up fine. Partitioned the new drive into 17G of ext3 and 2G of swap, deleted the old entries for the 100G from fstab and rebooted. Worked without a hitch. Shutdown and actually mounted the 20G in the case -- job done! Thanks JDD! -- David C. Rankin, J.D., P.E. Rankin Law Firm, PLLC 510 Ochiltree Street Nacogdoches, Texas 75961 Telephone: (936) 715-9333 Facsimile: (936) 715-9339 www.rankinlawfirm.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
jdd wrote:
David C. Rankin a écrit :
Listmates,
Critical situation. Server running 100G drive. The drive started to fail. Plenty of backups available and I was able to recovered all data to another server. Installed fresh 250G sata drives configured in hardware raid (sdb and sdc). Used the /swap from the failing drive (sda) for install purposes because I didn't was to put swap on the raid. Fresh 10.3 install replacing last mandrake server went fine. Server runs fine, just gives occasional /sda offline errors.
Problem, the old drive with /swap is sda, the /swap partition is sda5. The sda drive must be replaced, but how?
The question: "Can I just install a new sda, and boot and then partition a new /swap with yast, or, will the boot process freak out when /swap isn't available and cause the boot to fail?" I'm going to try it, but if someone else has done this and has advise to help avoid any "gotchas", I would definitely appreciate a heads up.
Thanks!
if you have enough ram, stop the swap (swapoff), unset the swap file in fstab (write #on beginning of the swap line).
no swap is really needed
that depends on how much memory the system has, and how much memory is consumed by the kernel and processes.
after that set any swap file on any disk (make swap, swap on) and re-enable the swap file
jdd
-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
David C. Rankin wrote:
Listmates,
1. I'ts not /swap, it's just swap. Unlike Windows, swap partitions do not need to be within any filesystem. More below.
Critical situation. Server running 100G drive. The drive started to fail. Plenty of backups available and I was able to recovered all data to another server. Installed fresh 250G sata drives configured in hardware raid (sdb and sdc). Used the /swap from the failing drive (sda) for install purposes because I didn't was to put swap on the raid. Fresh 10.3 install replacing last mandrake server went fine. Server runs fine, just gives occasional /sda offline errors.
Problem, the old drive with /swap is sda, the /swap partition is sda5. The sda drive must be replaced, but how?
The question: "Can I just install a new sda, and boot and then partition a new /swap with yast, or, will the boot process freak out
That depends on how much memory you have, and how much memory usage there is. But why not just use the Install/Repair CD and partition your new sda properly, and make sda5 type swap again?
when /swap isn't available and cause the boot to fail?" I'm going to try it, but if someone else has done this and has advise to help avoid any "gotchas", I would definitely appreciate a heads up.
-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Aaron Kulkis wrote:
But why not just use the Install/Repair CD and partition your new sda properly, and make sda5 type swap again?
Both sdb and sdc are fully partitioned in the hardware raid with: /dev/mapper/nvidia_haheegdb_part1 69972 15457 50902 24% /boot /dev/mapper/nvidia_haheegdb_part2 20641788 5114476 14478672 27% / /dev/mapper/nvidia_haheegdb_part3 219648612 12348660 196142408 6% /home Second, and most relevant, I was scared to death to mess with the fully functional array with a new install and temp fate with the partitioner screwing up the system. See: https://bugzilla.novell.com/show_bug.cgi?id=350573 "10.3 Promise pdc2036 Raid Install Crashes after Partitioning and Software Install" -- David C. Rankin, J.D., P.E. Rankin Law Firm, PLLC 510 Ochiltree Street Nacogdoches, Texas 75961 Telephone: (936) 715-9333 Facsimile: (936) 715-9339 www.rankinlawfirm.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
David C. Rankin wrote:
Aaron Kulkis wrote:
But why not just use the Install/Repair CD and partition your new sda properly, and make sda5 type swap again?
Both sdb and sdc are fully partitioned in the hardware raid with:
/dev/mapper/nvidia_haheegdb_part1 69972 15457 50902 24% /boot /dev/mapper/nvidia_haheegdb_part2 20641788 5114476 14478672 27% / /dev/mapper/nvidia_haheegdb_part3 219648612 12348660 196142408 6% /home
What does any of that have to do with sda?
Second, and most relevant, I was scared to death to mess with the fully functional array with a new install and temp fate with the partitioner screwing up the system. See:
if you properly size the partitions on sda, then the hardware raid should repopulate the other partitions.
https://bugzilla.novell.com/show_bug.cgi?id=350573 "10.3 Promise pdc2036 Raid Install Crashes after Partitioning and Software Install"
You're not installing software. I'm not going to ask you if you understand how RAID works, because even people who don't understand something often think that they do. So instead, I will ask you this: Why do you think that your RAID will not repopulate the other partitions on sda if they are part of the RAID structure? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Aaron Kulkis wrote:
David C. Rankin wrote:
Aaron Kulkis wrote:
But why not just use the Install/Repair CD and partition your new sda properly, and make sda5 type swap again?
Both sdb and sdc are fully partitioned in the hardware raid with:
/dev/mapper/nvidia_haheegdb_part1 69972 15457 50902 24% /boot /dev/mapper/nvidia_haheegdb_part2 20641788 5114476 14478672 27% / /dev/mapper/nvidia_haheegdb_part3 219648612 12348660 196142408 6% /home
What does any of that have to do with sda?
Aaron, sda was experiencing unrecoverable disc errors. It was dying and had to be replaced. That's what started this thread. The raid above doesn't contain swap and I didn't want to just remove sda and risk screwing up the raid created out of sdb and sdc by trying to repartition the discs in the raid array to try and shoehorn a swap partition on them. The downside is rebuilding the entire system again if something goes wrong and I don't have the time to jack with it. I'm not just talking about reloading the system, I'm talking about dhcp dyn dns postfix procmail mysql etc.. It may be easy and the partitioning may work, but it was a whole lot easier just to put a healthy sda in create swap and go... -- David C. Rankin, J.D., P.E. Rankin Law Firm, PLLC 510 Ochiltree Street Nacogdoches, Texas 75961 Telephone: (936) 715-9333 Facsimile: (936) 715-9339 www.rankinlawfirm.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
David C. Rankin wrote:
Aaron Kulkis wrote:
David C. Rankin wrote:
Aaron Kulkis wrote:
But why not just use the Install/Repair CD and partition your new sda properly, and make sda5 type swap again?
Both sdb and sdc are fully partitioned in the hardware raid with:
/dev/mapper/nvidia_haheegdb_part1 69972 15457 50902 24% /boot /dev/mapper/nvidia_haheegdb_part2 20641788 5114476 14478672 27% / /dev/mapper/nvidia_haheegdb_part3 219648612 12348660 196142408 6% /home
What does any of that have to do with sda?
Aaron,
sda was experiencing unrecoverable disc errors. It was dying and had to be replaced. That's what started this thread. The raid above doesn't contain swap and I didn't want to just remove sda and risk screwing up the raid created out of sdb and sdc by trying to repartition the discs in the raid array to try and shoehorn a swap partition on them.
The downside is rebuilding the entire system again if something goes wrong and I don't have the time to jack with it. I'm not just talking about reloading the system, I'm talking about dhcp dyn dns postfix procmail mysql etc.. It may be easy and the partitioning may work, but it was a whole lot easier just to put a healthy sda in create swap and go...
That's what I would do. Replace the failing disk, partition it properly, and resume computing. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Sat, 2008-01-05 at 00:38 -0500, Aaron Kulkis wrote:
That's what I would do.
Replace the failing disk, partition it properly, and resume computing.
Just a suggestion.... use lvm (except for /boot) If you want, you can grow any partion you want afterwards. Even on a life system! hw -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
participants (4)
-
Aaron Kulkis -
David C. Rankin -
Hans Witvliet -
jdd