You are not logged in.
Thanks for the reply! Yea, I just browsed through there and the only thing I'd question is that openssl update. I would in fact like to get any php 7.2 updates when their available which of course I wouldn't get if I disable that repo. I suppose it's entirely possible that something related to php 7.2 does in fact want or need that openssl version for some reason, though I would have expected it to install as a dependency of the php 7.2 stuff if that were the case(?).
I think the answer is no, but is there any way to block a specific package from an enabled repository?
EDIT: Just to note: I just testing and commenting out that repo and doing and update and dist-upgrade doesn't downgrade that openssl version. Also, after reading the home page at https://deb.sury.org/ it certainly appears that the repo is managed by someone who knows what they're doing.
Thanks!
Tom
I have a need to use PHP 7.2 on Devuan Jessie. I've been able to do so based on this:
https://tecadmin.net/install-php7-on-debian/
...which describes installing from this source:
https://packages.sury.org/php/README.txt
That all seems to have worked fine. One oddity was that we required php-pear and php-log which apparently needed to pull in php5-common php5-sqlite which had been removed, though that doesn't seem to have affected anything.
The one aspect of it that bothered me a little was that one of the updates that that repo included was openssl. Here are the upgradable packages after that an "apt update" with that source:
apt list --upgradable
Listing... Done
libgd3/unknown 2.2.5-5+0~20190119054525.2+jessie~1.gbp911a4a amd64 [upgradable from: 2.1.0-5+deb8u12]
libpcre3/unknown 2:8.42-1+0~20190203125145.5+jessie~1.gbp79d75d amd64 [upgradable from: 2:8.35-3.3+deb8u4]
openssl/unknown 1.1.1a-2~20190131152532.8+jessie amd64 [upgradable from: 1.0.1t-1+deb8u10]
php-pear/unknown 1:1.10.8+submodules+notgz-1+0~20190219091008.9+jessie~1.gbp1a209a all [upgradable from: 5.6.40+dfsg-0+deb8u1]As far as I could tell, it didn't appear that that openssl version was a dependency of any of the php stuff, and it didn't get installed until I expressly did an upgrade. Everything appears to be fine but I was curious as to that update, especially to something as important as openssl. I'm also curious as to those packages all reporting /unknown. Is that expected, and is there anything that can change that?
Thanks in advance if anyone knows something about any of that!
Tom
Figured I'd update this for anyone looking to do this. We got this working by adding the following modules to /etc/initramfs-tools/modules and then running update-initramfs:
virtio_rng
virtio_console
virtio_balloon
virtio_net
virtio_scsi
virtio_pci
virtio
virtio_ring
virtio_blk
rng_core
The reason we originally ran into trouble doing this was that we had omitted virtio_blk, so the device itself couldn't be accessed during the boot.
Tom
We use a Devuan Jessie VM (VMware) and were attempting to convert a copy of it for KVM. We converted it using qemu-img.exe under Windows using the flat VMware VMDK file. The issue we ran into was that initially on boot the main LVM volume group (containing the / file system) couldn't be read at all. From what I've read I'm fairly sure that this is because the initrd of the source VM probably doesn't have the virtio drivers, or something else that's required. In fact, a customer was apparently able to get it to boot by rebuilding the initrd, which supports that, though they weren't actually clear how they did that.
Our /etc/initramfs-tools/initramfs.conf currently has the default configuration with MODULES=most. Does anyone have any idea as to a configuration that would end up working for this after that conversion? I'm guessing that it might just be a matter of adding modules to /etc/initramfs-tools/modules, which was part of what we had to do to convert to Hyper-V.
Thanks!
Tom
We ran into a some strange issues with a customer that appear to have been caused by a stray dhclient process. I've been able to figure out why that process was running and how to prevent that. However I'm trying to determine if that process may have been causing other issues we had. This is with Devuan 1.0 Jessie by the way. Also note that we're not using network manager.
The symptom we had was that some services were getting arbitrarily stopped. Not crashing, but seeming stopped intentionally. This included some of our own services and on one occasion, mysql. In the case of the latter, the mysql log clearly showed a clean shutdown that was not initiated by anyone as far as we know.
We deliver our product on a VM configured for DHCP by default. In this case there was either no DHCP server available, or possibly it refused to give us a lease, so the network start timed out. That in and of itself wasn't an issue. Our install process configures a static IP etc as specified by the customer. What I discovered was that we were writing a new /etc/network/interfaces file before stopping the network...which prevented the stop process from knowing that there was a dhclient process to kill. We've since changed that.
While that process was running as it turned out, the /var/log/syslog file showed evidence that we were getting repeatedly hit with DHCPDISCOVER requests from that processes...at least one a second actually. While I'm not 100% clear on this, I believe there were signs in /var/log/messages that this may have been causing the interface to go up and down occasionally. Unfortunately we've yet to find out whether the services have properly kept running since we killed that process.
So my big question is this: Could that process repeatedly trying to get an IP for that statically configured adapter somehow cause services to stop like that?...possibly due to the LSB dependencies or the like? We've certainly never seen anything remotely like this before.
Thanks!
Tom
Thanks! Yea that's what I thought. In our case these will always be headless servers in customer environments where X11Forwarding won't be an option. So the only means of cleaning up (which appears to only involve the unused linux-image-* packages in this case) would be with a careful apt-get purge.
Thanks for the replies. I was clearly misunderstanding dist-upgrade quite a bit.
Tom
I was a bit confused as to what dist-upgrade does and was under the impression it did much more major updating than it does. I just tried that on a copy of the same VM I updated as per my last post. It did in fact install the updated kernel (related to the installed linux-image-amd64 meta package). Note however that it didn't automatically uninstall the old one as you described (and frankly I wouldn't want it to):
dpkg -l | grep linux-image
ii linux-image-3.16.0-5-amd64 3.16.51-3+deb8u1 amd64 Linux 3.16 for 64-bit PCs
ii linux-image-3.16.0-6-amd64 3.16.56-1+deb8u1 amd64 Linux 3.16 for 64-bit PCs
ii linux-image-amd64 3.16+63+deb8u2 amd64 Linux for 64-bit PCs (meta-package)I'm not using backports or the like. I'm assuming that's why it just upgraded to a newer 3.16 kernel?
Thanks! I was somewhat mistaken as to what dist-upgrade does.
Tom
I'm reviving an old topic here to clarify something regarding these kernel upgrades (within the same major version). Yesterday I upgraded my kernel on a Jessie install and here's what I ran into. I just wanted to make sure this is expected:
There was clearly a new kernel available for the linux-image-amd64 meta package. When I tried to get this via "apt-get upgrade" here's what I got:
apt-get upgrade
Reading package lists... Done
Building dependency tree
Reading state information... Done
Calculating upgrade... Done
The following packages have been kept back:
linux-image-amd64
0 upgraded, 0 newly installed, 0 to remove and 1 not upgraded.What I finally figured out was that I could get the new kernel using this instead:
apt-get install --only-upgrade linux-image-amd64
Reading package lists... Done
Building dependency tree
Reading state information... Done
The following extra packages will be installed:
linux-image-3.16.0-6-amd64
Suggested packages:
linux-doc-3.16 debian-kernel-handbook
The following NEW packages will be installed:
linux-image-3.16.0-6-amd64
The following packages will be upgraded:
linux-image-amd64
1 upgraded, 1 newly installed, 0 to remove and 0 not upgraded.
Need to get 34.5 MB of archives.
After this operation, 168 MB of additional disk space will be used.
Do you want to continue? [Y/n] That worked and appears to have run update-grub as well. When I rebooted I was running that new version. Apparently at that point if I wanted to remove the unused kernel, I could then do:
apt-get purge linux-image-3.16.0-5-amd64
update-grubI'm assuming the "apt-get upgrade" may not work simply because upgrade implies installing the new version and removing the existing(?) which is obviously unsafe for kernels. Is that correct?
Thanks!
Tom
Well I did the update just now, reloading Synaptic just showed the Linux image metapackage and one other as upgradable, not the kernel itself, but when I upgraded the two packages it installed the new kernel but in addition to the older kernel, so now have two. I assume the older can be deleted but just wondering why it didn't do it as an upgrade instead?
If I understand what you're saying, I believe the same happened to me, and I'd assume that's the desired behavior. I have the linux-image-amd64 meta package installed, and after adding the jessie-security source the upgrade pulled in linux-image-3.16.0-5-amd64 (with the new 3.16.51-3+deb8u1 version) in addition to my existing linux-image-3.16.0-4-amd64, making the newer the default in grub.
If the new kernel were to cause any issues, I'd be able to boot to the old one using the arrow keys in the grub menu. Since it was OK, after rebooting, I uninstalled the old one, and did an update-grub. If that's what you're referring to I'd expect that as a safeguard.
Those are all valid jessie repositories, pick and choose. Some may not have a contrib and non-free components.
Thanks! I think the fact that I used the expert non-graphical install may be how I ended up with just the jessie source. I've now added the security source. For our purposes I think that will do for now, as adding either the updates or the backports actually doesn't cause anything new to install anyway. Sorry for sending the thread a bit off topic too.
I have to say that I did find your sources.list quite strange-looking... ^__^; Are you sure no one "tinkered" with it?
That wasn't the entire file, but it was in fact the only uncommented source in the file. The only change I made to the original one from the install was to comment out the line for the install CD ISO, as I was packaging it as a VM. The file also has the sources for the deb-src commented.
I can tell you for sure that jessie-security wasn't in there.
EDIT: Are backports and updates also supposed to be in there? Those were never in mine either. I used the expert non-graphical install. I'm wondering if there wasn't something in that that I selected incorrectly, though I don't recall that.
Tom
It looks like you are missing jessie-security... Try adding this:
deb http://us.mirror.devuan.org/merged/ jessie-security main non-free contrib
That doesn't do it either. What it gets me is even more confusing:
apt list --upgradable | grep linux
WARNING: apt does not have a stable CLI interface yet. Use with caution in scripts.
linux-image-3.16.0-4-amd64/stable 3.16.51-2 amd64 [upgradable from: 3.16.43-2+deb8u2]
linux-image-amd64/jessie-security 3.16+63+deb8u1 amd64 [upgradable from: 3.16+63]I have no clue what that second one even is.
I do in fact have the linux-image-amd64 meta package installed. Nothing I do seems to find that linux-image-3.16.0-5-amd64, but rather newer versions of linux-image-3.16.0-4-amd64. Earlier I tried adding jessie-updates and that didn't work either and was even more confusing. That wanted to pull in a version of linux-image-3.16.0-4-amd64 that was 3.16.51-3, but NOT the 3.16.51-3+deb8u1 that actually has the fix. Insane.
My company is sort of really dying for this one. I hope someone has suggestions because this is getting more confusing with everything I try.
SOLVED: OK...Now I think I get it. That new version of the linux-image-amd64 meta package from jessie-security (3.16+63+deb8u1) actually ends up pulling in linux-image-3.16.0-5-amd64 with the 3.16.51-3+deb8u1 version.
That was confusing. Is there some reason that the default sources.list from the install wouldn't include that by default? I can also see we were missing a number of updates because of that.
Tom
tlathm wrote:Am I missing something here or am I maybe just hitting a mirror that hasn't synced yet or something?
I assume you did run "apt-get update", so I guess you are indeed using a mirror not updated yet...?
Definitely. I just updated to 3.16.51-2 and now it tells me everything's up to date. Really seems odd. This is all I have enabled in /etc/apt/sources.list:
deb http://us.mirror.devuan.org/merged/ jessie main non-free contribIs that correct?
Tom
It looks like it is already available!
# apt-get -s install --only-upgrade linux-image-amd64 Reading package lists... Done Building dependency tree Reading state information... Done The following extra packages will be installed: firmware-linux-free irqbalance libnuma1 linux-image-3.16.0-5-amd64 [snip] Inst linux-image-3.16.0-5-amd64 (3.16.51-3+deb8u1 None:1.0/jessie-security [amd64]) [snip]
This is odd. I currently have 3.16.43-2+deb8u2 installed. I've done "apt update", however "apt list --upgradable" only shows 3.16.51-2 and not 3.16.51-3+deb8u1:
apt list --upgradable | grep linux
linux-image-3.16.0-4-amd64/stable 3.16.51-2 amd64 [upgradable from: 3.16.43-2+deb8u2]Am I missing something here or am I maybe just hitting a mirror that hasn't synced yet or something?
EDIT: Now I'm even more confused. Bear with me as I'm fairly new to apt etc, but when I run the exact command you did I get this:
apt-get -s install --only-upgrade linux-image-amd64
Reading package lists... Done
Building dependency tree
Reading state information... Done
linux-image-amd64 is already the newest version.
0 upgraded, 0 newly installed, 0 to remove and 74 not upgraded.Why does that not at least show the 3.16.51-2? Totally lost now.
Thanks!
Tom
I just checked and apparently Debian now has a Meltdown fix for jessie (3.16.51-3+deb8u1 apparently):
https://security-tracker.debian.org/tra … -2017-5754
Does this mean we might have an update for jessie in the not to distant future? Thanks!
Tom
I think I'm clear on what we're going to try here. Looking at the existing steps we took to use our old CentOS 6 VM in hyperv, we didn't install any services at all. All we did was to ensure that the initramfs included hv_utils, hv_vmbus, hv_storvsc, and hv_netvsc. So I think all we need to do is to add those to
/etc/initramfs-tools/modules and run "update-initramfs -u". Hopefully that's all we need.
Tom
We're hoping to convert a copy of our headless Devuan VM (jessie) to be used in HyperV...yuk...wouldn't if we didn't have to. I'm hoping we don't run into issues.
What exactly is that status of the current hyperv-daemons package in Devuan? I was reading some Debian jessie centric stuff around this like this:
https://oitibs.com/install-hyper-v-lis-on-debian-8/
There's a lot there that I can't quite make sense of. I see that there's a hyperv-daemons in the standard repo, but this is all it contains:
dpkg-query -L hyperv-daemons
/.
/lib
/lib/systemd
/lib/systemd/system
/lib/systemd/system/hyperv-daemons.hv-vss-daemon.service
/lib/systemd/system/hyperv-daemons.hv-kvp-daemon.service
/lib/systemd/system/hyperv-daemons.hv-fcopy-daemon.service
/usr
/usr/share
/usr/share/doc
/usr/share/doc/hyperv-daemons
/usr/share/doc/hyperv-daemons/changelog.Debian.gz
/usr/share/doc/hyperv-daemons/copyright
/usr/share/doc/hyperv-daemons/README.Debian
/usr/sbin
/usr/sbin/hv_fcopy_daemon
/usr/sbin/hv_kvp_daemon
/usr/sbin/hv_vss_daemonNotably nothing around init scripts if they're needed for example.
It appears there are some related modules for the current kernel:
find /lib/modules |grep hv
/lib/modules/3.16.0-4-amd64/kernel/drivers/scsi/hv_storvsc.ko
/lib/modules/3.16.0-4-amd64/kernel/drivers/hv
/lib/modules/3.16.0-4-amd64/kernel/drivers/hv/hv_balloon.ko
/lib/modules/3.16.0-4-amd64/kernel/drivers/hv/hv_utils.ko
/lib/modules/3.16.0-4-amd64/kernel/drivers/hv/hv_vmbus.ko
/lib/modules/3.16.0-4-amd64/kernel/drivers/net/hyperv/hv_netvsc.koI do see however that the notes in the above link around /etc/initramfs-tools/modules and updating the initramfs mentions hv_blkvsc, which I don't see there.
I'm also unclear as to how much of any of that is actually required to run under HyperV.
Thanks in advance for any info!
Tom
I agree - old grub was so much more understandable...
Totally. Under Gentoo I moved to Syslinux when I saw how grub 2 worked. Not a fan.
That's what I figured. I don't have Synaptic as this is a headless server. Thanks!
Tom
I have Jessie with the linux-image-amd64 meta package installed. Most of what I read about kernel upgrades seems to be around doing dist-upgrade. I'm a little confused as far as ongoing kernel upgrades that might occur within the same major version.
If (and I assume there are) occasional kernel upgrades available for Jessie, do they just get installed via api-get update and apt-get upgrade? If so, does it leave the existing kernel installed etc. Also...and this is probably related...does it automatically run grub-update if I'm using grub?
Thanks!
Tom
Thanks! I'll look into that. It appears it might be overkill given the situation. Manually adding that library does get this CLI working, and frankly this CLI is for a very rare corner case that out product arguably shouldn't even be supporting at the moment.
Thanks again.
Tom
I actually was able to get the cli working by manually installing the library extracted from the downloaded deb file as described here:
https://preterhuman.net/forum/index.php?topic=98.0
...but that seems more than a little ugly. Still wondering if there's a better way. I noticed looking at the package details:
https://packages.debian.org/jessie/libstdc%2B%2B5
...that it appears to require an i386 libc6 >= 2.3, and I see that my install appears to have libc6-i386 2.19. Maybe that's related(??), though this seems to work.
Tom
I'm trying to get an old CLI utility working in Jessie that requires a 32bit libstdc++5. Am I missing something or is that not available?:
apt install libstdc++5:i386
Reading package lists... Done
Building dependency tree
Reading state information... Done
E: Unable to locate package libstdc++5
E: Couldn't find any package by regex 'libstdc++5'Thanks in advance!
Tom
of course, no need to update initrd. My mistake. I am sorry
No problem. Thanks! I'm not sure I we'd actually ever need that kernel command line, as I'm getting the eth0 name anyway, but I've added it to the grub config be safe.
Tom
you can set
GRUB_CMDLINE_LINUX="net.ifnames=0 biosdevname=0"
then update grub/initrdthen whatever will happen to init you should always get consistent eth0
remember though to etit interfaces.
Thanks! Just to clarify, being a bit new to this: I assume that the change there is in the /etc/default/grub file, and when you say update grub you mean update-grub, and as far as the initrd, is that "update-initramfs -u"?
I'm too used to Gentoo where I'm using syslinux and no initrd at all. Thanks!
EDIT: Thinking about this more: Why would this require any change to the initrd? Wouldn't simply changing /etc/default/grub and running update-grub be sufficient? Thanks.
Tom