Re: [GLLUG] Debian Live pxe boot
On Tue, 27 Jun 2023 at 10:49, Ken Smith via GLLUG mailto:gllug@mailman.lug.org.uk>> wrote: Ken Smith via GLLUG wrote: > Hi All, > > I wonder if someone with Debian and pxe understanding might point me > in the right direction. > > I have a Rocky server with dhcp/tftp/httpd that hosts Rocky and Fedora > as pxe bootable images. It works fine. > > I'm trying to do similar for Debian 12 but not quite getting the right > incantations. > > The vmlinuz and initrd load just fine and i get the error "Unable to > find live system filesystem on the network" > > I pointed the root filesystem at filesystem.squashfs. Is the standard > vmlinuz/initrd able to map the rest of the system via httpd?? > > Any tips welcome > > Thanks > > Ken > > I'm making progress. I added dhcp ethdevice=eno1 to the parameters in the pxe configuration and now one of my PC's finds the filesystem.squashfs and attempts to boot. It freezes at "Starting Gnome Display Manager" I tried another system that has more memory and it fails with a message that it can't find eno1. Interesting. :-) Ken John Hearns via GLLUG wrote: Ethernet device naming is a minefield, even with the predictable names (yeah I know.. do some studying) I set biosdevname=1 in the kernel parameters when building HPC clusters In the end I got it working by leaving out the "ethdevice=" statement and letting the booting system work it out. :-) Thanks Ken -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean. -- GLLUG mailing list GLLUG@mailman.lug.org.uk https://mailman.lug.org.uk/mailman/listinfo/gllug
Re: [GLLUG] Debian Live pxe boot
Ken Smith via GLLUG wrote: Hi All, I wonder if someone with Debian and pxe understanding might point me in the right direction. I have a Rocky server with dhcp/tftp/httpd that hosts Rocky and Fedora as pxe bootable images. It works fine. I'm trying to do similar for Debian 12 but not quite getting the right incantations. The vmlinuz and initrd load just fine and i get the error "Unable to find live system filesystem on the network" I pointed the root filesystem at filesystem.squashfs. Is the standard vmlinuz/initrd able to map the rest of the system via httpd?? Any tips welcome Thanks Ken I'm making progress. I added dhcp ethdevice=eno1 to the parameters in the pxe configuration and now one of my PC's finds the filesystem.squashfs and attempts to boot. It freezes at "Starting Gnome Display Manager" I tried another system that has more memory and it fails with a message that it can't find eno1. Interesting. :-) Ken -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean. -- GLLUG mailing list GLLUG@mailman.lug.org.uk https://mailman.lug.org.uk/mailman/listinfo/gllug
[GLLUG] Debian Live pxe boot
Hi All, I wonder if someone with Debian and pxe understanding might point me in the right direction. I have a Rocky server with dhcp/tftp/httpd that hosts Rocky and Fedora as pxe bootable images. It works fine. I'm trying to do similar for Debian 12 but not quite getting the right incantations. The vmlinuz and initrd load just fine and i get the error "Unable to find live system filesystem on the network" I pointed the root filesystem at filesystem.squashfs. Is the standard vmlinuz/initrd able to map the rest of the system via httpd?? Any tips welcome Thanks Ken -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean. -- GLLUG mailing list GLLUG@mailman.lug.org.uk https://mailman.lug.org.uk/mailman/listinfo/gllug
Re: [GLLUG] sendmail puzzle
Marco van Beek via GLLUG wrote: On 10/10/2022 22:54, Ken Smith via GLLUG wrote: I'm trying to sort out a Rocky 8.5 server that has sendmail installed. (Please don't go on a diversion about how I should tell the owner to dump sendmail and switch to exim or postfix - save that for another thread please. ) I'm pretty good with sendmail but this problem has me a bit foxed. I'd value some suggestions of where to look as I think I'm not seeing the wood for the trees. It will send from addresses in the local network, without auth, because I have "connect:192.168.123 relay" in the access file - that being the local LAN. I've tested sasl auth and that is authenticating. Using telnet from an IP off their LAN (over a VPN) I can connect using TLS (openssl s_client etc etc) and authenticate (perl -MMIME::Base64 etc etc) it accepts my credentials and then if I try to send a message I get "Relaying denied. IP name lookup failed [my local ip]" The same happens with a test using Thunderbird. If I do the same test from the host that sendmail is on, it works fine. Also if I do the same test from another host on the same LAN it works fine. Somehow its complaining about the source IP in authenticated sessions outside the LAN range. In the test from my local LAN (172.16.0.x), over a VPN, the remote dns can't resolve the reverse dns of my LAN. I've done a similar test with another sendmail site and remote auth is working fine. Maybe sendmail is doing reverse DNS when it shouldn't be. Suggestions most welcome Thanks Ken Hi, It might be the difference between a missing entry in a zone file, and a missing zone file. Maybe it is the lookup mechanism that fails, rather than it checking the IP address itself. It might be another rule set that is trying to do a reverse lookup (eg hostname), and it barfs out at that point. Maybe increase the logging verbosity and check the logs again? Cheers, Marco Thank you - Not sure where my error was but, probably a typo on my part. I reconfigured it from the ground up using the template config files I've kept from other setups and its working fine now. Didn't touch any thing to do with DNS or /etc/hosts. All fixed. Yay :-) Ken -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean. -- GLLUG mailing list GLLUG@mailman.lug.org.uk https://mailman.lug.org.uk/mailman/listinfo/gllug
[GLLUG] sendmail puzzle
I'm trying to sort out a Rocky 8.5 server that has sendmail installed. (Please don't go on a diversion about how I should tell the owner to dump sendmail and switch to exim or postfix - save that for another thread please. ) I'm pretty good with sendmail but this problem has me a bit foxed. I'd value some suggestions of where to look as I think I'm not seeing the wood for the trees. It will send from addresses in the local network, without auth, because I have "connect:192.168.123 relay" in the access file - that being the local LAN. I've tested sasl auth and that is authenticating. Using telnet from an IP off their LAN (over a VPN) I can connect using TLS (openssl s_client etc etc) and authenticate (perl -MMIME::Base64 etc etc) it accepts my credentials and then if I try to send a message I get "Relaying denied. IP name lookup failed [my local ip]" The same happens with a test using Thunderbird. If I do the same test from the host that sendmail is on, it works fine. Also if I do the same test from another host on the same LAN it works fine. Somehow its complaining about the source IP in authenticated sessions outside the LAN range. In the test from my local LAN (172.16.0.x), over a VPN, the remote dns can't resolve the reverse dns of my LAN. I've done a similar test with another sendmail site and remote auth is working fine. Maybe sendmail is doing reverse DNS when it shouldn't be. Suggestions most welcome Thanks Ken -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean. -- GLLUG mailing list GLLUG@mailman.lug.org.uk https://mailman.lug.org.uk/mailman/listinfo/gllug
Re: [GLLUG] Rocky Linux on HP ML350G6 - Motherboard failure??
Chris Bell wrote: On Sunday, 15 May 2022 17:32:50 BST you wrote: Chris Bell via GLLUG wrote: On Sunday, 15 May 2022 13:37:10 BST Ken Smith via GLLUG wrote: Any pearls of wisdom either about my question or about recovering a miscreant ML350 motherboard that won't boot anything or go into the setup pages. Thanks Ken Hello Ken, I had a very lightly loaded ML150 running last year when suddenly it totally stopped Did yours recover? No permanent damage, it was just the same as an ordinary re-start After finding how to engage the secondary BIOS, I reflashed the primary BIOS and the machine has been running the 5.4 Kernel for over 24 hours now. I don't think the mysterious crash was to do with the 5.4 Kernel after all. Motherboard saved from the recyclers! Ken -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean. -- GLLUG mailing list GLLUG@mailman.lug.org.uk https://mailman.lug.org.uk/mailman/listinfo/gllug
Re: [GLLUG] Rocky Linux on HP ML350G6 - Motherboard failure??
Chris Bell via GLLUG wrote: On Sunday, 15 May 2022 13:37:10 BST Ken Smith via GLLUG wrote: Any pearls of wisdom either about my question or about recovering a miscreant ML350 motherboard that won't boot anything or go into the setup pages. Thanks Ken Hello Ken, I had a very lightly loaded ML150 running last year when suddenly it totally stopped (but without any permanent damage) when the ambient room temperature reached its maximum working temperature. Did yours recover? -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean. -- GLLUG mailing list GLLUG@mailman.lug.org.uk https://mailman.lug.org.uk/mailman/listinfo/gllug
[GLLUG] Rocky Linux on HP ML350G6 - Motherboard failure??
Hi All, This might seem a bit of an off the wall question. Would running the 5.4 Kernel stress a system such that it could cause a motherboard failure? I can't think how but. More details — I’m running (or was running) various KVM VM’s as well as MythTV server on the host machine with Rocky Linux 8.5 (Kernel 4.18) with a couple of DVB PCIe capture cards. I have found that there is an anomaly in the stock DVB driver for those cards, in the 4.18 kernel causing the DVB card to stop responding in some situations. I loaded 5.4 Kernel from ELRepo, that is supposedly built from the standard config. Although it does not have the DVB media driver stack, I built the current DVB Media stack from source. The anomaly is supposed to be fixed in the current version of DVB media drivers. All was working fine for a few hours and this, I hope, is complete co-incidence, the machine crashed and it now appears that the BIOS is corrupt. It will show some of the Post messages but the options to enter BIOS Setup (RBSU in HP speak) don’t work and it won’t boot anything. Neither the Raid array, a USB stick or from PXE on the network. I’ve tried swapping memory and stripping the machine down to the bare minimum to start up and resetting it to factory default. No luck. I have a replacement motherboard on order. My health monitoring of this machine with Zabbix did not log any significant anomalies in the last few moments of uptime, no excess temperatures, power consumption or processor load etc. ILO, which still works, does not show any errors or alerts either. But one doubt I have is that this happened just 3 hours after I started running the 5.4 kernel. It seems a bit far fetched, but is that part of the picture here or just pure coincidence? I realise I could simply replace the DVB driver stack in the stock kernel by compiling it up from source. That is an option too. Any pearls of wisdom either about my question or about recovering a miscreant ML350 motherboard that won't boot anything or go into the setup pages. Thanks Ken -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean. -- GLLUG mailing list GLLUG@mailman.lug.org.uk https://mailman.lug.org.uk/mailman/listinfo/gllug
Re: [GLLUG] LVM / KVM Question
Andy Smith via GLLUG wrote: Hi, On Tue, Feb 15, 2022 at 10:19:32AM +, Ken Smith via GLLUG wrote: I'm contemplating a system where I pass a LVM LV through to a KVM VM. On the host I'm thinking that I could do a backup of the LV using a LVM snapshot. What am I missing, is this a mad idea? Many people do this. However: - Backed up data from the snapshot won't contain working data from within the memory of the processes in the VM, so if you restored from it, it would be like coming up after a power loss. - Lots of other KVM disk backends can do snapshots so LVM isn't required to do this. Cheers, Andy Thank you Andy, I agree a backup 'in flight' like this may have compromised integrity. I'm also looking at QCOW snapshots. :-) Ken -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean. -- GLLUG mailing list GLLUG@mailman.lug.org.uk https://mailman.lug.org.uk/mailman/listinfo/gllug
[GLLUG] LVM / KVM Question
Hi All, I'm contemplating a system where I pass a LVM LV through to a KVM VM. On the host I'm thinking that I could do a backup of the LV using a LVM snapshot. What am I missing, is this a mad idea? Thanks Ken -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean. -- GLLUG mailing list GLLUG@mailman.lug.org.uk https://mailman.lug.org.uk/mailman/listinfo/gllug
Re: [GLLUG] Multi-Boot Puzzle - a bit OT
John Winters via GLLUG wrote: On 04/10/2020 18:21, James Roberts via GLLUG wrote: [snip] I've done this a lot in the past and now I just don't. Disks are cheaper than my time, so I am prepared to multi-boot, but not from one disk. [snip] IIWM I'd acquire another disk or two and put everything on its own disk. Seconded. Reading the original query I felt like I'd dropped through a time warp and gone about 20 years into the past. I wasn't aware that people did such convoluted installations any more. Disks are ridiculously cheap these days. Give everything its own disk. I'd take a safe image/copy of that 1.5 tib disk as my first action Also seconded. John Hi Guys, Well this machine all started out simple.. Anyway. I've come to the same conclusion, separate disks, in my reading around the topic today. At some point I might even virtualise the legacy Svr 2K3 system. I think I may also have stubbed my toe on a XP/Svr 2K3 issue related to booting from other than sector 63. Thank you for your time - much appreciated. :-) Ken -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean. -- GLLUG mailing list GLLUG@mailman.lug.org.uk https://mailman.lug.org.uk/mailman/listinfo/gllug
[GLLUG] Multi-Boot Puzzle - a bit OT
Hi All, I have a machine I use for Music recording. It has an old Fedora 13 install and it also boots Server 2003 and Win 7. Its high time it was upgraded to something current and my plan is to add Fedora 32 and Win 10. I'd like to keep the old installs as they host music software I'd like to keep available. It had a 1.5 TB disk that I have copied using dd to a 2TB disk. Forgetting about the 512 / 4096 block size issue I've ended up with some misaligned partitions and the performance is terrible as a result. To try to correct this I created correctly aligned partitions/correct size on another 2TB disk and copied the partitions to it. Fedora 13 boots fine but Grub chainloading the windows installs fails. I suspect the Windows installs are missing their boot sector loaders, but my attempts with Windows boot sector repair tools are failing. It looks like Win 7 has put its boot files on the partition for Server 2003 as there is a /boot in there. I suspect this partition being NTFS is a factor in this problem. But the original 1.5TB disk is the same and it works. I've reinstated the FC13 systems Grub boot loader. I guess I could put Fedora 32 and Win 10 on another disk and use Grub2 from FC32 to also boot the systems on the original 1.5TB disk. Thank for reading this far. Any suggestions/recomendations folks. Ken -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean. -- GLLUG mailing list GLLUG@mailman.lug.org.uk https://mailman.lug.org.uk/mailman/listinfo/gllug
Re: [GLLUG] KVM Performance
James Courtier-Dutton via GLLUG wrote: On Fri, 12 Jun 2020 at 08:55, Tim Woodall via GLLUG wrote: On Tue, 9 Jun 2020, Ken Smith via GLLUG wrote: Hi All, While in lockdown I decided to do some performance testing on KVM. Interesting, I've been looking into this myself trying to improve performance/reduce cpu usage. Hi, My experience with KVM is that it is pretty good, performance wise, with block devices. E.g. qcow2 or actual block devices. KVM appears very slow for USB devices. Some of your tests appear to include USB in the path. USB devices are also very slow when used with WINE. (Running windows app on Linux) Kind Regards James Not in my tests. Mine had two SATA 7200rpm 3TB disks. -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean. -- GLLUG mailing list GLLUG@mailman.lug.org.uk https://mailman.lug.org.uk/mailman/listinfo/gllug
Re: [GLLUG] KVM Performance
Mike Brodbelt via GLLUG wrote: On 09/06/2020 22:52, Ken Smith via GLLUG wrote: I can' t help thinking that there must be something more behind a drop from 60MB/s to 15MB/s in write performance Can't speak to your exact bottleneck here, but I think I'd try and eliminate your disk hardware from the game for a start, to see what happens there. Try creating a null block device (https://www.kernel.org/doc/html/latest/block/null_blk.html), and then benchmark writes to that on the host and the guest. See what that does as a baseline Mike Interesting - thanks for the tip. I measured 16.3 MB/s write to a null block device on one of the guests. Conclusion - that's as fast as that hardware can manage. :-) Ken -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean. -- GLLUG mailing list GLLUG@mailman.lug.org.uk https://mailman.lug.org.uk/mailman/listinfo/gllug
Re: [GLLUG] KVM Performance
Martin A. Brooks via GLLUG wrote: The more levels of indirection you have between any 2 components in a system the slower stuff will move. Indeed. So wouldn't passing a block device from the host through to the guest minimise the 'components' that are 'in the way' I can' t help thinking that there must be something more behind a drop from 60MB/s to 15MB/s in write performance I'd value your thoughts :-) Ken -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean. -- GLLUG mailing list GLLUG@mailman.lug.org.uk https://mailman.lug.org.uk/mailman/listinfo/gllug
[GLLUG] KVM Performance
Hi All, While in lockdown I decided to do some performance testing on KVM. I had believed that passing a block device through to a guest rather than using a QCOW2 file would get better performance. I wanted to see whether that was true and indeed whether using iSCSI storage was any better/worse. My test hardware is quite modest and this may adversely have affected what I measured. The processor is a Intel Core2 6300 @ 1.86GHz with VT-X support. It shows 3733 Bogomips at startup. There's 8GB RAM and an Intel 82801HB SATA controller on a Gigabyte MB. The disks are two 3TB SATA 7200RPM set up with a Raid 1 LVM Ext3 partition as well as other non-Raid partitions to use to test. I used Fedora 32 as the KVM host and my testing was with Centos 8 as a guest. On the host I got 60MB/s write and 143 MB/s read on Raid1/LVM/Ext3. I wrote/read 10GB files using dd. 10Gb so as to overflow any memory based caching. Without LVM that changed to 80 MB/s write and 149 MB/s read. I tried all kinds of VM setups. Normal QCOW2, pass though of block devices Raid/LVM and Non-Raid/LVM. I consistently got around 14.5 MB/s write and 16.5 MB/s read. Similar figures with iSCSI operating from both file based devices and block devices on the same host. The best I got by tweaking the performance settings in KVM was a modest improvement to 15 MB/s write and 17 MB/s read. As a reference point I did a test on a configuration that has Centos 6 on Hyper-V on an HP ML350 with SATA 7200 rpm disks. I appreciate that's much more capable hardware, although SATA rather than SAS, but I measured 176 MB/s write and 331 MB/s read. That system is using a file on the underlying NTFS file system to provide a block device to the Centos 6 VM. I also tried booting the C8 guest via iSCSI on a Centos6 Laptop, which worked fine on a 1G network. I measured 16.8 MB/s write and 23.1 MB/s read that way. I noticed an increase in processor load while running my DD tests, although I didn't take any actual measurements. What to conclude? Is the hardware just not fast enough? Are newer processors better at abstracting the VM guests with less performance impact? What am I missing?? Any thoughts from virtualisation experts here most welcome. Thanks Ken -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean. -- GLLUG mailing list GLLUG@mailman.lug.org.uk https://mailman.lug.org.uk/mailman/listinfo/gllug