1

My setup is as follows

/dev/sdb/ /dev/sdc/ /dev/sdd/ /dev/sdf/

I have the following 4 1.5TB drives in a mdadm raid 5 array. One of the disks (sdb) died on me leading the array to start up running in degraded mode.

No problem I thought as this is the whole reason I setup a raid 5 array in the first place. I had a spare blank 1.5tb disk to replace the dead sdb disk, so I formatted it and added it into the array. However this is where the problem really starts.

When recovering the array after adding a fresh disk the recovery fails every single time, though not always at the same percentage. After recovery fails sdc is marked as a failed device and also removed from the array though it can be forced to be added again.

Here's some of the outputs from syslog

Jan 5 01:16:28 serverlol kernel: [11303.917452] md/raid:md0: Disk failure on sdc1, disabling device. Jan 5 01:16:28 serverlol mdadm[3345]: Fail event detected on md device /dev/md0, component device /dev/sdc1

Here's the output of syslog when the failure occurs

Jan 5 01:16:24 serverlol kernel: [11300.853422] end_request: I/O error, dev sdc, sector 693768801 Jan 5 01:16:24 serverlol kernel: [11300.853426] md/raid:md0: read error not correctable (sector 693766752 on sdc1). Jan 5 01:16:24 serverlol kernel: [11300.853429] md/raid:md0: read error not correctable (sector 693766760 on sdc1). Jan 5 01:16:24 serverlol kernel: [11300.853432] md/raid:md0: read error not correctable (sector 693766768 on sdc1). Jan 5 01:16:24 serverlol kernel: [11300.853434] md/raid:md0: read error not correctable (sector 693766776 on sdc1). Jan 5 01:16:24 serverlol kernel: [11300.853436] md/raid:md0: read error not correctable (sector 693766784 on sdc1). Jan 5 01:16:24 serverlol kernel: [11300.853438] md/raid:md0: read error not correctable (sector 693766792 on sdc1). Jan 5 01:16:24 serverlol kernel: [11300.853441] md/raid:md0: read error not correctable (sector 693766800 on sdc1). Jan 5 01:16:24 serverlol kernel: [11300.853443] md/raid:md0: read error not correctable (sector 693766808 on sdc1). Jan 5 01:16:24 serverlol kernel: [11300.853446] md/raid:md0: read error not correctable (sector 693766816 on sdc1). Jan 5 01:16:24 serverlol kernel: [11300.853448] md/raid:md0: read error not correctable (sector 693766824 on sdc1).

I have run smartctl on /dev/sdc and I see that there are definitely some errors, I'm running the same tests on the other drives in the array but it takes a very long time to complete on drives this big.

Ok so here are the recovery steps I have taken so far, I have ordered an external 4tb hard disk onto which I intend to back up as much data as I possibly can from the degraded array. This is because the array can still be mounted and browsed whilst in degraded mode. Though of course presumably this will not be the case for all of the data contained on it. I have currently stopped mdadm and unmounted the array whilst I wait for the 4tb backup disk to arrive.

My current plan is to use the tool "safecopy" to recover as much data as I can to the external hard drive. Is this the best utility to copy the data from my degraded /dev/md0 device?

Is this my best course of action to recover as much of my data as I can? Or is there any thing more I can do to fix the problem?

I have a replacement disk for the initial failed drive (sdb) but I do not have any more 1.5 tb drives to replace the other one being marked as failed (sdc). This is why I just want to recover as much data as I can to the external drive, after that I will probably just build an entirely new array since these 1.5TB drives are pretty old.

0

You must log in to answer this question.

Browse other questions tagged .