2020. 2. 20. 09:22ㆍ카테고리 없음
I have a HP server which has a P800 SCSI controller. The CCISS kernel module is required to run the server.On a server I have running 16.04 LTS, the module lives in /lib/modules/4.4.0-103-generic/kernel/drivers/block/cciss.ko and is by default part of the initrd. After upgrading a server that was also running 16.04 LTS to 18.04 LTS (using do-release-upgrade -d) the system has been rendered unbootable since grub doesn't have all modules loaded to mount the rootfs.Is the module in some other package which needs to be installed?Is Canonical aware of this? (as in, was this a conscious decision?) If so, what's the reasoning behind removing the module?
If removing the module was intended, why is the package 'cciss-vol-status' still offered?This server is an Amanda backup server, so this is more than a mere annoyance - if I can't find a solution I'll have to revert it to 16.04 LTS - but it also rules out Ubuntu on half my network. The reason however for the requirement of upgrade is that Amanda had started to exhibit regression errors which weren't being fixed in 16.04 LTS.Anyone with knowledge, please advise. Admin@thjalfi: $ man cciss. + Smart Array P400 + Smart Array P400i + Smart Array P600 + Smart Array P700m + Smart Array P800Apparently this is a Linus-.up, not a Canonical one since the alledgedly supported boards don't work 'as advertised'.So these machines are at EOL where Linux is concerned (unless I find a magic way to fix this) - I thought it would be a coldday in Hell when I found a machine that had a longer life with Windows than with Linux.
Hpacucli
And this is the second machine I findthat Linusx has 'given up on'. I guess that explains the sleet, hail, snow and rain outside my window.