Bug 156081
| Summary: | Unable to set faulty drives in linear mode | ||
|---|---|---|---|
| Product: | [Fedora] Fedora | Reporter: | Paolo Campegiani <p.campegiani> |
| Component: | mdadm | Assignee: | Doug Ledford <dledford> |
| Status: | CLOSED NOTABUG | QA Contact: | |
| Severity: | medium | Docs Contact: | |
| Priority: | medium | ||
| Version: | 4 | CC: | mattdm |
| Target Milestone: | --- | ||
| Target Release: | --- | ||
| Hardware: | i386 | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Fixed In Version: | Doc Type: | Bug Fix | |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2007-07-02 15:16:19 UTC | Type: | --- |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Embargoed: | |||
[Mass switching of bugs filed against FC4T3 to fc4test3 from test3] This report targets the FC3 or FC4 products, which have now been EOL'd. Could you please check that it still applies to a current Fedora release, and either update the target product or close it ? Thanks. Fedora Core 3 and Fedora Core 4 are no longer supported. If you could retest this issue on a current release or on the latest development / test version, we would appreciate that. Otherwise, this bug will be marked as CANTFIX one month from now. Thanks for your help and for your patience. Closing this bug as NOTABUG. The issue is that non-redundant raid modes can't be set faulty. The entire concept of tracking faulty disks is only useful if you have some means of recovering from the fault. If you don't, then the errors are simply propagated back up to the higher levels of the OS, such as the file system code, where they are handled. The ioctl that mdadm uses to set a device faulty hooks into the raid subsystem at the same point that a real error hooks into it. Since the non-redundant arrays just pass errors up to the higher layers, the mdadm call succeeds in trigering the code, but the code essentially does nothing. |
Description of problem: If you define a linear RAID, a faulty device does not compromise the entire array, nor /proc/mdstat gives a clue about that. How reproducible: Always. # uname -a Linux linux5.confor.it 2.6.11-1.14_FC3 #1 Thu Apr 7 19:23:49 EDT 2005 i686 i686 i386 GNU/Linux # mdadm --version mdadm - v1.6.0 - 4 June 2004 # mdadm --create --verbose /dev/md0 --chunk 16 --level raid0 --raid-devices=4 /dev/hda5 /dev/hda6 /dev/hda7 /dev/hda8 So I want a linear array composed of some partitions of /dev/hda (I know, but it's a test). mdadm: /dev/hda5 appears to contain an ext2fs file system size=457536K mtime=Wed Apr 27 13:18:41 2005 mdadm: /dev/hda5 appears to be part of a raid array: level=0 devices=4 ctime=Wed Apr 27 13:20:34 2005 mdadm: /dev/hda6 appears to be part of a raid array: level=0 devices=4 ctime=Wed Apr 27 13:20:34 2005 mdadm: /dev/hda7 appears to be part of a raid array: level=0 devices=4 ctime=Wed Apr 27 13:20:34 2005 mdadm: /dev/hda8 appears to contain an ext2fs file system size=-76546048K mtime=Wed Apr 27 13:18:41 2005 mdadm: /dev/hda8 appears to be part of a raid array: level=0 devices=4 ctime=Wed Apr 27 13:20:34 2005 Continue creating array? y mdadm: array /dev/md0 started. # cat /proc/mdstat Personalities : [raid0] md0 : active raid0 hda8[3] hda7[2] hda6[1] hda5[0] 610048 blocks 16k chunks unused devices: <none> And this seems to me a bit strange, as /proc/mdstat does not contain the [UUUU] string. Nevertheless, the array is up and running: # lsmod | grep raid raid0 8129 1 # mkfs -t ext3 /dev/md0 mke2fs 1.35 (28-Feb-2004) max_blocks 156172288, rsv_groups = 4766, rsv_gdb = 37 [...] This filesystem will be automatically checked every 33 mounts or 180 days, whichever comes first. Use tune2fs -c or -i to override. # mount /dev/md0 /raid # ls /raid lost+found I test it by copying some files in, and it works. # mdadm --manage /dev/md0 --set-faulty /dev/hda6 mdadm: set /dev/hda6 faulty in /dev/md0 # ls /raid [...] same as above # cat /proc/mdstat Personalities : [raid0] md0 : active raid0 hda8[3] hda7[2] hda6[1] hda5[0] 610048 blocks 16k chunks Shouldn't /proc/mdstat tell that the linear RAID array is gone? I can still copy files to it! Actual results: RAID array continues to work. Expected results: RAID array should be gone. Additional info: if I build a RAID-5 array with the same partitions, I've the [UUUU] in /proc/mdstat, and faulty-ing one array component is reported in /proc/mdstat (and I can remove it and hot-add it, seeing reconstruction progress and so on).