On 7/30/18 10:25 AM, Theodore Y. Ts'o wrote: > On Mon, Jul 30, 2018 at 12:44:46PM +0800, Patrick Dung wrote: >> The problem usually appeared when I did not use the hard drive for a while. >> It happened a few times in the past. >> >> When I perform fsck today, it does not appear. >> I had checked the SATA hard drive with smartmontools. It passed the >> long test and I did not found any problem. >> >> After searching the web. I found Cisco WebEX Node SPA have very >> similar error message that I had encountered. >> https://www.cisco.com/c/en/us/td/docs/interfaces_modules/shared_port_adapters/install_upgrade/ASR1000/asr_sip_spa_hw/ASRtrbl.pdf >> Please check page 7-8 or search for "Inode 7" in the document. > > I've looked at the Cisco documentation, and what you've cited appears > to be an example of how to fix a corrupted file system. It appears to > me to be merely an example, not an acknowledgement this kind of > corruption regularly happens on the Cisco WebEx device (if it were, > customers would probably refuse to buy it, since it means a system > administrator would be regularly needing to do this kind of manual > intervention on what is *supposed* to be an appliance sort of device.) > > In any case, I can't really tell you much more. We do a lot of > extensive regression testing, and this is not a corruption I've seen > before. A lot of people use ext4, and you are the first person who > has reported this particular problem. > > This is why I suspected that it might be a hardware problem. It's > possible it is a kernel bug, and it could be anything, from an ext4 > bug, to a device driver bug, to an LVM or MD or bcache bug if you are > using any of those components. You didn't report the kernel version > to me, and you didn't tell me if this is a distro kernel. If it is a > distro kernel, I'd suggest reporting it to the distribution, since > they might be able to tell you if anyone else with that distro kernel > has reported a problem similar to yours. FYI, Patrick reported it on the Fedora bugzilla with kernel: 4.17.3-200.fc28.x86_64 e2fsprogs-1.44.2-0.fc28.x86_64 Trolling around the RH bugzilla I do see other instances of a corrupt resize inodes, but it was always in combination with large swath of corruption surrounding it as well, i.e. in one case due to a faulty storage driver for an SDXC card.... -Eric
