NETGEAR is aware of a growing number of phone and online scams. To learn how to stay safe click here.
Forum Discussion
lexa2
Mar 23, 2012Aspirant
[SOLVED]ReadyNas Duo v1 on-disk raid format vs. linux mdraid
Hello 2all,
My relative had been happy using ReadyNas Duo for about two years till recent month when suddenly things went out wrong without prior warning. He had received a mail from NAS telling about the failure of one of the installed hard drives, and after trying to reboot NAS shortly afterwards it gone bonkers - no answers to requests over the network (pings, web access, e.t.c), no new mails from NAS, nothing at all. As main goal is to try to recover all the data possible from the HDDs, he proceed with turning off the box, ejecting HDDs and bringing them to a person he thinks who's able to do the recovery duties - for this case this person turned out to be me. Actually, I'm a *nix system engineer specializing in linux and FreeBSD system administration and programming, so in general it shouldn't be too tough for me to at least diagnose the case.
From what I see, disks themselves are not physycalli broken (at least - not totally broken), I've been able to do full raw data backup using dd tool. I've been told that the disks had been set up in a non-raid fashion (i.e., separate volume for every HDD), and inspecting partition table of grabbed images seems to confirm it: I can see three partitions arrangement on each of images, 2048MB, 512MB and "rest of the disk". All three partitions have "linux mdraid" partition type, 2GB and 512MB volumes seems to be raid1 (mirror) ones and I thinks they are most likely to be NAS "/" and swap partitions. Third volume is a linux mdraid raid0 1 drive volume, which makes sense for "separate disks" scenario. My problem is that standard linux mdadm toolchain can't handle this volumes as they report that the "superblock checksum" is bad. My guess is that NETGEAR had used a modified mdraid kernel module and mdadm toolchain inside their ReadyNAS so the disks ejected from it can't be directly used by standard linux installation.
So, here are my questions:
a) As this modifications had been done over GPL'ed code - where may I fetch the sources for the modified kernel module and mdadm utility?
b) In case I would re-create the mdraid volume using standard linux mdadm toolchain - would the resulting array use the same storage data ranges as ReadyNAS modified mdadm raid0 use?
Thanks in advance for answers.
P.S. Please don't post something like "here are contacts of a company that offers recovery services in your country", it'd be pointless and would only result in waste of time. My questions are well-determined and I don't see much of a problem answering them for a techincally-equipped person from ReadyNAS team in case NETGEAR internal rules allow that.
My relative had been happy using ReadyNas Duo for about two years till recent month when suddenly things went out wrong without prior warning. He had received a mail from NAS telling about the failure of one of the installed hard drives, and after trying to reboot NAS shortly afterwards it gone bonkers - no answers to requests over the network (pings, web access, e.t.c), no new mails from NAS, nothing at all. As main goal is to try to recover all the data possible from the HDDs, he proceed with turning off the box, ejecting HDDs and bringing them to a person he thinks who's able to do the recovery duties - for this case this person turned out to be me. Actually, I'm a *nix system engineer specializing in linux and FreeBSD system administration and programming, so in general it shouldn't be too tough for me to at least diagnose the case.
From what I see, disks themselves are not physycalli broken (at least - not totally broken), I've been able to do full raw data backup using dd tool. I've been told that the disks had been set up in a non-raid fashion (i.e., separate volume for every HDD), and inspecting partition table of grabbed images seems to confirm it: I can see three partitions arrangement on each of images, 2048MB, 512MB and "rest of the disk". All three partitions have "linux mdraid" partition type, 2GB and 512MB volumes seems to be raid1 (mirror) ones and I thinks they are most likely to be NAS "/" and swap partitions. Third volume is a linux mdraid raid0 1 drive volume, which makes sense for "separate disks" scenario. My problem is that standard linux mdadm toolchain can't handle this volumes as they report that the "superblock checksum" is bad. My guess is that NETGEAR had used a modified mdraid kernel module and mdadm toolchain inside their ReadyNAS so the disks ejected from it can't be directly used by standard linux installation.
So, here are my questions:
a) As this modifications had been done over GPL'ed code - where may I fetch the sources for the modified kernel module and mdadm utility?
b) In case I would re-create the mdraid volume using standard linux mdadm toolchain - would the resulting array use the same storage data ranges as ReadyNAS modified mdadm raid0 use?
Thanks in advance for answers.
P.S. Please don't post something like "here are contacts of a company that offers recovery services in your country", it'd be pointless and would only result in waste of time. My questions are well-determined and I don't see much of a problem answering them for a techincally-equipped person from ReadyNAS team in case NETGEAR internal rules allow that.
7 Replies
- mdgm-ntgrNETGEAR Employee RetiredO.K. This is a Duo v1 not a Duo v2. Duo v2 was released Nov 13, 2011.
It's better to use dd_rescue rather than dd to clone the disks as dd_rescue is designed to clone failing disks.
There are instructions for recovering data from a ReadyNAS array. See Mounting Sparc-based ReadyNAS Drives in x86-based Linux
However these are for X-RAID not Flex-RAID and you're using Flex-RAID.
You'll still need fuseext2 and lvm2.
The Sparc systems e.g. Duo v1 use a custom mdconfig. See http://www.readynas.com/forum/viewtopic.php?p=263334#p263334 - lexa2AspirantThanks for provided links, going to head on and read info available there.
As for ReadyNas model - I had asked my relative about the exact information that's written at the sticker on the bottom of the box and he told me that it is: "ReadyNAS Duo RND2000 v2". Don't know how should I interpret the "v2" there - is it a Duo v1 really? Seems like the info on the sticker is a bit misleading. - StephenBGuru - Experienced UserThe sticker is misleading. Netgear also uses v2, v2, v3 for minor hardware updates, regrettably they didn't think much about that when they named the brand-new NAS line.
Since you are dealing with a 2 year old product it absolutely is a V1. There is a comprehensive guide on this topic: http://www.rnasguide.com/2012/01/09/how ... -or-nv-v2/ - mdgm-ntgrNETGEAR Employee RetiredYes that is misleading.
Long before the Duo v2 was developed there was a minor hardware change to the Duo v1 (I think probably one forced due to availability of parts). So NetGear put that sticker on units to indicate the difference. Probably useful for NetGear to be able to quickly determine what hardware they're dealing with when troubleshooting defective units returned via RMAs. Have a read of e.g. How to tell whether I have a Duo v1 or Duo v2 or NV+ v1 or NV+ v2 which lists several ways you can tell the difference between a Duo v1 and Duo v2.
Another thing worth noting is the Duo has a 3 year warranty (if purchased brand new). So if you put some new disks from the http://www.readynas.com/hard_disk_hcl in the Duo and still have problems with it you should open a tech support case. If support determines the unit is defective then it could be replaced with a refurbished unit via RMA.
Edit: beaten. - lexa2AspirantThanks once again for exhaustive answers. So, yes, accordingly to the info in the "guide on topic", this box is really Duo v1 and not Duo v2. As for warranty matters - hadn't had a chance to investigate if the box itself is really bricked yet. The only things I had been provided with so far are HDDs as the info recovery was the main thing that my relative had been interested in. I wonder why hadn't he been using mirrored raid for "such a valued stuff" he had been storing on the NAS, but that's an offtopic question really which can lead to a lengthy discussion about people carelessness.
As for RMA - the box had been bought brand-new, but chances are good that I wouldn't be able to RMA it due to the specifics we have here in Russia when dealing with warranties. As far as I have heard, NETGEAR haven't got dedicated office in Moscow, so the only place we would be able to query about the warranty is a retail computer store where this box had been bought - and RMA success rate for such cases is far less than 100%, as long as my experience tells.
Anyways, thanks once again for quick answers and help, would try to proceed with data recovery on this weekend. - mdgm-ntgrNETGEAR Employee RetiredIt could just be a bad disk or a filesystem issue or something like one of those. Most likely the NAS chassis is still O.K.
- lexa2AspirantJust FYI: I was able to successfully recover all the data from both HDDs without having to resort to compiling/using custom mdadm on freshly installed CentOS 6 using fuse-ext2 v.0.0.7 compiled from sources.
Tricky part was to force lvm2 into detecting HDD partitions as PVs - default LVM2 configuration filters out from detection partitions which have "linux raid autodetect" type and/or contain mdraid superblock signatures. To turn this behavior off one would have to edit the /etc/lvm/lvm.conf file, find a line saying "md_component_detection = 1" and change it to "md_component_detection = 0". Having this done (make sure your existing mdraid/lvm2 setup wouldn't be messed up with this change - you should know if this is possible and how to avoid it as long as you're geeky enough to use lvm2 :-D) proceed with the usual vgscan/vgchange -ay magic and with luck you would end up with /dev/c/c and /dev/d/d lvm2 block devices become available for mount with fuse-ext2.
Feel free to PM me (or reach by e-mail at mooroon2 (at sign goes here) mail [here is dot] ru) if more details/advices would be needed concerning the "recovering data from ReadyNas Duo using ordinary linux pc" process.
I would like to thank once again all the posters to this topic, provided hints were invaluable.
Related Content
NETGEAR Academy
Boost your skills with the Netgear Academy - Get trained, certified and stay ahead with the latest Netgear technology!
Join Us!