Re: [Users] Fedora 10 i386/x86_64 default/minimal OS Templates released to contrib
Default AND minimal, 32 AND 64 bit? Thanks, Scott, this is excellent! -- Gregor Mosheh / Greg AllensworthBS, A+, Network+, Security+, Server+ System Administrator, Lead Programmer HostGIS development & hosting services, http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] Setting UB params where / how ?
Adem wrote: Where can I set the UB parameters like kmemsize etc. (cf. below)? That's in the VPS's config file, e.g. /etc/vz/conf/123.conf It looks like your config file didn't get created for some reason; I presume you used "vzctl create" in the usual fashion? -- Gregor Mosheh / Greg AllensworthBS, A+, Network+, Security+, Server+ System Administrator, Lead Programmer HostGIS development & hosting services, http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] unable to migrate / checkpoint
Scott Dowdle wrote: I think there is a quota file on the filesystem that can be moved > too... but if you don't, it'll just make a new one... FYI: The quota files are in /var/vzquota However, each quota file has the pathname inside it, so if you rename the quota file it is no longer valid. And if you get clever and modify the binary file, it breaks the checksum and invalidates it anyway. Either way, you will have to sit through vzquota init and start, which can be VERY tedious if you deal with 100-200 GB VPSs like I do. But you may want to delete the old quota file, in case you reuse the ID# later. -- Gregor Mosheh / Greg AllensworthBS, A+, Network+, Security+, Server+ System Administrator, Lead Programmer HostGIS development & hosting services, http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] kmemsize totals?
Scott Dowdle wrote: Just out of curiousity, what values does vzsplit with -n 32 recommend? 25 being more realistic, I used -n 25 and it suggests 200M kmemsize. Well that answers my question about a paltry 64M apiece! -- Gregor Mosheh / Greg AllensworthBS, A+, Network+, Security+, Server+ System Administrator, Lead Programmer HostGIS development & hosting services, http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
[Users] kmemsize totals?
Second to privvmpages, kmemsize seems to be the limit our users hit most often. Setting it to 64M seems sufficient when 32M isn't. My burning question: What limit exists on the total kmemsize for all VEs on a HN? Can I have 32 machines at 64M apiece, for 2 GB of kmemsize? -- Gregor Mosheh / Greg AllensworthBS, A+, Network+, Security+, Server+ System Administrator, Lead Programmer HostGIS development & hosting services, http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
[Users] Intermittent ARP failures using veth
Hiya, all. I have 3 HNs running a total of some 30 VEs. There are three specific VEs which have a problem which seems ARP-related. Any ideas on it would be much appreciated. Symptom: Every now and then, sometimes every minute and sometimes not for a whole hour, the VE becomes inaccessible via IP for some 10-30 seconds. During this time, other hosts in the network also can't ping the host. The "arp" command shows (incomplete) for the host's entry. Temporary fix: Using arpsend on the HN "/usr/sbin/arpsend -U -i $ip -c1 bond0" fixes this issue until the client expires its ARP entry. I have a cronjob to run this every minute, but even that isn't enough. Other IP and routing info: * There are 5 IP blocks, /27 and /28 in size. IPs from all blocks are arbitrarily distributed around the machines. * We have 2 GigE switches. Each HN has dual GigE NICs, and uses Linux bonding. The 2 NICs go to the 2 switches, for fault tolerance. * At the border we have a router which we don't control. Traffic between IP blocks, even if destined for the local network, is double-transited. The problem only affects these three and none of the others, and its affecting only these specific three has been retained even as VEs are moved between HNs. As such, I have ruled out bad cables or switch ports, overloading of the hardware, system load, and differences in the HN's OS and sysctl params. All VEs are created using the same script; the only diffs in their VZ config would be auto-generated MACs and veths. I've been over these pages, and couldn't find information to help: http://wiki.openvz.org/Multiple_network_interfaces_and_ARP_flux http://wiki.openvz.org/Virtual_Ethernet_device. Any thoughts on troubleshooting this? Any further information I should provide? -- Gregor Mosheh / Greg Allensworth, BS, A+ System Administrator HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] ovzkernel-xen ?
Robert Nelson wrote: So I have a Xen PV OpenVZ DomU that hosts a container for each Linux Distro and Version I need to build. So why the complexity of the hybrid setup, rather than a separate DomU for each OS and distro? I have my reasons for loving OpenVZ, and am curious about reasons for your situation. -- Gregor Mosheh / Greg AllensworthBS, A+, Network+, Security+, Server+ System Administrator, Lead Programmer HostGIS development & hosting services, http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: AW: [Users] ovzkernel-xen ?
Dietmar Maurer wrote: On Proxmox VE (pve.proxmox.com), we combine KVM and openvz. That way we are able to run fully virtualized guests (i.e. windows) and openvz CTs. Oh, that is just killer. And you have that phat GUI and a bare-metal installer? That is SO neat. I'll want to hit you up off-list, or on the Proxmox mailing list, to find out more about the nesting model and interoperability etc. Way cool. -- Gregor Mosheh / Greg Allensworth, BS, A+ System Administrator HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] Re: ovzkernel-xen ?
Benny Amorsen wrote: department gets the ability to migrate, consolidate on physical servers, do snapshot backups and all the other nifty things ... then you put OpenVZ on it just to make administration easier > for you (no more multi-purpose machines) without OpenVZ for separating functions, and VMWare under it to make migration and recovery a snap. That's very snazzy. Thanks for sharing. -- Gregor Mosheh / Greg Allensworth, BS, A+ System Administrator HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] ovzkernel-xen ?
Josip Rodin wrote: On Fri, Aug 22, 2008 at 04:22:01PM -0600, Gregor Mosheh wrote: My question is why you'd want to use both. What particular need or deployment scenario would require Xen and OpenVZ together? Well, that's... a bit unimaginative :| I'm trying to get an idea of what creative needs and solutions are being handled with nested virtualization, not trawling for insults. But thanks anyway. :) Here's a quick example: you have one big machine but more than one group of users Interesting. So you hand out a Xen instance, expecting the owner of the DomU to create OpenVZ instances within it? That's kinda clever, I guess, if there's some practical use for it. I'm dying to know the real-life scenario underlying this, why you don't hand someone either a Xen or a OpenVZ, but a Xen so they can create their own OpenVZs. Do you have multiple IT departments, for example, working on different products, and you want them to self-administer the creation of VEs? Do you have some 128-CPU cluster running Xen and you sell Xen Doms so they can resell OpenVZ VEs? -- Gregor Mosheh / Greg Allensworth, BS, A+ System Administrator HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] ovzkernel-xen ?
Josip Rodin wrote: Yes, there is a fundamental difference in how the two systems work, RTFM :) Why, thank you. And here I thought they were the same thing. :P My question is why you'd want to use both. What particular need or deployment scenario would require Xen and OpenVZ together? -- Gregor Mosheh / Greg Allensworth, BS, A+ System Administrator HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] ovzkernel-xen ?
Having not used Xen but being very interested in all forms of virtualization and how folks are using it, I'm curious. What does one do with a kernel that does both OpenVZ and Xen? Do you run multiple Doms and a few of them use OpenVZ? Is there a reason that neither OpenVZ nor Xen do exactly what you want by themselves? -- Gregor Mosheh / Greg Allensworth, BS, A+ System Administrator HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] sysctl to hide VE processes from HN ps?
Kir Kolyshkin wrote: # sysctl kernel.pid_ns_hide_child=1 Processes are hidden for all containers started _after_ you have changed the setting. A, I get it. Too bad it can't change the view more dynamically. I'd love to be able to toggle the view of HN or HN+VPSs. But I'll likely not be making that patch; I'll be patient. :) -- Gregor Mosheh / Greg AllensworthBS, A+, Network+, Security+, Server+ System Administrator, Lead Programmer HostGIS development & hosting services, http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] 2.6.24 stable
Roberto Mello wrote: Is 2.6.24 considered stable yet? When I installed it three weeks ago, it was considered development. I've been using it with DRBD, NIC bonding, and a 3ware 9650 controller for the last three weeks and so far so good! -- Gregor Mosheh / Greg AllensworthBS, A+, Network+, Security+, Server+ System Administrator, Lead Programmer HostGIS development & hosting services, http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
[Users] sysctl to hide VE processes from HN ps?
I was glad to hear about bug 511. I am running 2.6.24-ovz005 and was surprised to see that I had this great new feature. But it doesn't seem to be working. Maybe I'm doing it wrong? # sysctl kernel.pid_ns_hide_child=1 A 'ps ax' or 'ls /proc' shows the same thing regardless of this setting: VPSs' PostgreSQL and httpd and so on. -- Gregor Mosheh / Greg Allensworth, BS, A+ System Administrator HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] DRBD version?
Kirill Korotaev wrote: AFAIK, 8.0 is a stable branch. though sure update to 8.0.12 is required long ago... But in reality, I was simply asked to include exactly this versions by the developers working with DRBD. If they believe it must/can be updated to 8.2 branch - why not? Huh; how odd. Well, I'll run some tests on 8.2 and let you know what I find: compiling, functionality, etc. -- Gregor Mosheh / Greg Allensworth, BS, A+ System Administrator HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
[Users] DRBD version?
Why does OpenVZ (for Fedora anyway) come with such an old version of DRBD (8.0.7) instead of the newer 8.2 series? Is there any reason I shouldn't or couldn't upgrade DRBD? -- Gregor Mosheh / Greg Allensworth, BS, A+ System Administrator HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] New to the List, Great Project!
JR Richardson wrote: Hi All, Just wanted to send a test email to the list. I'm sincere, if not original, in expressing my "Holy cow, this rocks" gratitude to the developers. It's rock-solid, fast as can be, and highly flexible since beancounters can be adjusted on the fly. We've been using OpenVZ for over a year now, are hosting over 30+ servers, and are now planning the next stage of our hosting service's infrastructure, still based on OpenVZ. Thanks to OpenVZ, we're able to achieve high availability of almost-equal quality (seconds of downtime during HA failover), better backups and easier management, and about 1/3 the cost, versus hard quotes about that other virtualization product (name rhymes with "CMHair"). It also reinforces our company's stated commitment to open-source solutions, which puts a warm fuzzy in my belly too. "Thanks" and "We couldn't have done it without you" barely cover it. :) -- Gregor Mosheh / Greg Allensworth, BS, A+ System Administrator HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
[Users] ARP attacks in OpenVZ?
Hey there. I am curious to what degree VEs (using veth, to be exact) can engage in ARP attacks, e.g. ARP poisoning, ARP spoofing, and so on. I'm asking not only about VEs within a HN but a VE spoofing a IP on another HN on the same LAN. I see that the HN performs proxy ARP, but are VEs actually prevented from generating ARP traffic? I know too that veth routing is very stringent, e.g. veth10.0 is the only route to a given IP, so I suspect that ARP spoofing even if possible would be ineffective. Any other thoughts on the topic? -- Gregor Mosheh / Greg Allensworth, BS, A+ System Administrator HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] Skipping quota at start and doing it later?
Gregor wrote:
Sure, but we DO want quotas to be enforced. I just don't want to wait
for recalculating disk usage at startup, when I know that the
usage hasn't changed since the power went out.
Kir Kolyshkin wrote:
It was suggested earlier in this thread that if you will comment out the
following line
need_restart="${need_restart} ${veid}"
in /etc/init.d/vz you get the exact behavior you ask for. Can you
confirm (or deny) this?
I did not try that mechanism. I didn't see "--nocheck" in the vzquota
man page, and when running it manually during the initial steps of the
problem, the flag had no effect (quota checking still ran). As such, I
didn't accept that "vzquota on --nocheck" would accomplish my goal, even
if the later step of restarting were skipped.
I already had in place a change to the init script, which runs "vzquota
off" before running "vzquota on" in order to repair the quota file and
then come back up.
It goes something like this, and I'd be glad to provide the complete script.
# fastboot means to turn quota off,
# effectively repairing the quota file or doing nothing
if QUOTA and VZFASTBOOT; then vzquota off $VEID ; fi
# now turn quota on as usual
if QUOTA; vzquota on $VEID ; fi
I'm sure some folks actually like the current behavior, of their VPSs
starting, then restarting over the course of an hour to recalculate disk
usage, but this hack worked great for my needs.
--
Gregor Mosheh / Greg Allensworth, BS, A+
System Administrator
HostGIS cartographic development & hosting services
http://www.HostGIS.com/
"Remember that no one cares if you can back up,
only if you can restore." - AMANDA
___
Users mailing list
Users@openvz.org
https://openvz.org/mailman/listinfo/users
Re: [Users] Skipping quota at start and doing it later?
Kir Kolyshkin wrote: I guess that if you can live w/o disk quotas you can disable this entirely and forget about the problem you have. Option to set is DISK_QUOTA=no (in /etc/vz/vz.conf) Sure, but we DO want quotas to be enforced. I just don't want to wait 90 minutes for recalculating disk usage at startup, when I know that the usage hasn't changed since the power went out. -- Gregor Mosheh / Greg Allensworth, BS, A+ System Administrator HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] Skipping quota at start and doing it later?
Matt Ayres wrote:
What you want is not possible with any existing options. You could
possibly edit the vz init script and removed the needs_restart line.
A, okay. Then I guess my hack to /etc/init.d/vz is the best way to
accomplish what I want. We already have a policy of never bypassing the
tmpfs, so the quota stats before a power outage are good enough for our
needs.
FYI, the fix I have applied is probably not the best, but here it is for
posterity:
start_ves()
{
local veid
local velist
local msg
local need_restart
need_restart=""
cd ${CONFIG_DIR} || return
velist=`grep -l '^ONBOOT=yes\|^ONBOOT=\"yes\"' [0-9]*.conf
2>/dev/null |
sed -e 's/.conf//g' | sort -n`
cd - >/dev/null
sysctl -q -w net.ipv4.route.src_check=0
for veid in $velist; do
[ "${veid}" = "0" ] && continue
__echo "Starting VE ${veid}: "
if [ "x${VZFASTBOOT}" = "xyes" -a "x${DISK_QUOTA}" =
"xyes" ]; then
$VZQUOTA off ${veid} >/dev/null 2>&1
fi
msg=`$VZCTL start ${veid} 2>&1`
print_result "$msg"
done
}
--
Gregor Mosheh / Greg Allensworth, BS, A+
System Administrator
HostGIS cartographic development & hosting services
http://www.HostGIS.com/
"Remember that no one cares if you can back up,
only if you can restore." - AMANDA
___
Users mailing list
Users@openvz.org
https://openvz.org/mailman/listinfo/users
Re: [Users] Skipping quota at start and doing it later?
Matt Ayres wrote: VZFASTBOOT is the option. /etc/vz/vz.conf is the main Virtuozzo config, not a VPS/Container config file Ah; my mistake. I misread you and the script. This sets a flag that this VPS needs to be restarted normally later on in the script. The normal restart will cause the quota to be fully checked. I assure you this is the option you want enabled. Hm, I'll take your word for it, though it doesn't sound right. I don't want them to be restarted with a full quota recalculation - I don't want quota recalculation at all. These fellas have 250 GB quotas and are using most of it, so the recalculation takes 60+ minutes per VPS. -- Gregor Mosheh / Greg Allensworth, BS, A+ System Administrator HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] Skipping quota at start and doing it later?
Matt Ayres wrote: All you have to do is set VZFASTBOOT=yes in /etc/vz/vz.conf and VPS's will be started without quota and then will be restarted to calculate the quota after all have been started. I think I see where we're misconnecting. You say that FASTBOOT will cause all VPSs to start, then those with fastboot will be re-started later so their quota info can run? What I'd really like it to skip the quota recalculation entirely, achieve the same effect I currently do with vzquota off (repairing quota file without recalculating) and vzquota on (turning quota on). -- Gregor Mosheh / Greg Allensworth, BS, A+ System Administrator HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] Skipping quota at start and doing it later?
Matt Ayres wrote: Which version of the vzctl RPM do you have installed? On my vzctl-3.0.22-1 I see the following in /etc/init.d/vz: My versions: vzctl-3.0.22-1 vzctl-lib-3.0.22-1 ovzkernel-2.6.18-53.1.13.el5.028stab053.10 vzquota-3.0.11-1 I have the same paragraph in the init script. What I read is: If fastboot and disk_quota are both yes, make sure that quota isn't already running, then run vzquota on. In our case, neither FASTBOOT nor DISK_QUOTA are set in the conf files, yet vzquota runs anyway (normally good, but after a power failure very time consuming). The --nocheck option is not mentioned in the man pages nor does it seem effective. What am I missing here? My hack was to add a "vzquota off" line right above the "vzquota on" line. It's effective, but not as great as a "real" solution. -- Gregor Mosheh / Greg Allensworth, BS, A+ System Administrator HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] Skipping quota at start and doing it later?
Matt Ayres wrote: All you have to do is set VZFASTBOOT=yes in /etc/vz/vz.conf and VPS's will be started without quota and then will be restarted to calculate the quota after all have been started. Interesting. I saw FASTBOOT in the /etc/init.d/vz script and that's not what I saw. I'll give it a try, though. Thanks for the note. -- Gregor Mosheh / Greg Allensworth, BS, A+ System Administrator HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
[Users] Skipping quota at start and doing it later?
Hey guys. Some of our VPSs take well over an hour to start up after an unexpected shutdown, thanks to vzquota. Reading the man page, it seems that vzquota cannot be skipped and then done later once the services are up. Am I correct about that, or IS there a way to skip quota checking until after the VPS has started? -- Gregor Mosheh / Greg Allensworth, BS, A+ System Administrator HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] How to reset the failcnt?
Jorge Fuertes wrote: Can we reset the failcnt to see the new errors? Nope, except by stopping and restarting the VE. This is by design, the reason being that some programs would expect the number to increase and would be confused if it were to decrease or be 0. Personally, I think it would be spiffy to be able to reset the failcounts, and I don't know what specific programs they have in mind which would be so confused by the failcounts being zeroed. -- Gregor Mosheh / Greg Allensworth, BS, A+ System Administrator HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] terminal resizing
Dmitry V. Levin wrote: See http://bugzilla.openvz.org/show_bug.cgi?id=683 Consider updating to vzctl>=3.0.19. That's the stuff! Thanks a zillion. -- Gregor Mosheh / Greg Allensworth System Administrator, HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
[Users] terminal resizing
Hey, all. I have a recurring issue, and finally thought to ask about it. I use PuTTY to login to my HN. I then use vzctl enter to enter a VPS. As long as I am in the VPS, I cannot resize my putty window and have the VPS's terminal follow it. e.g. if I make the window taller while running top, top will not extend to fill the space below; if I'm using vi, vi will think that the terminal is the old size and I get weird effects trying to edit anything. Any thoughts on why this happens, or how I can work around it? I've tried experimenting with stty and the TERM variable, but so far nothing helps. -- Gregor Mosheh / Greg Allensworth System Administrator, HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
[Users] a simple utility to execute something on all VEs
http://wiki.openvz.org/ExecuteInAllVEs Simple but useful. If anybody has further enhancements, that'd be spiffy. -- Gregor Mosheh / Greg Allensworth System Administrator, HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] Cloning and permissions
Steve Wray wrote: tar cf - * | ( cd ../2 ; tar xvf - ) That doesn't need a -p option? ie I tend to use 'tar --numeric-owner -cf' as well, just in case. I've never used the -p flag but it's always preserved the numerics just fine. That's Slackware and Fedora, probably others. Couldn't hurt to use it, though. -- Gregor Mosheh / Greg Allensworth System Administrator, HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] Cloning and permissions
Peter Machell wrote: vzctl stop xx cp -R /vz/private/xx /vz/private/xxx cp -R /etc/vz/conf/xx.conf /etc/vz/conf/xxx.conf vzctl start xxx To have cp preserve permissions, use the -p flag. I use tar instead of cp, for situation like this. Tar, unlike cp, is smart enough to handle symbolic links, permissions, ownerships, and sparse files. Use this age-old Unix Jedi trick: mkdir /home/vz/private/2 cd /home/vz/private/1 tar cf - * | ( cd ../2 ; tar xvf - ) -- Gregor Mosheh / Greg Allensworth System Administrator, HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] strange problem with nagios nrpe server
Kirill Korotaev wrote: Just for the history/other users the resolution of the problem Steve had: OpenVZ was installed on XFS WOW, good work Kirill. That must have been a gnarly one to figure out, I never even thought of the filesystem type combined with a bug in NRPE. Hopefully, now Steve can join the ranks of us highly satisfied OpenVZ users. ) -- Gregor Mosheh / Greg Allensworth System Administrator, HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] Help disk qouta error
Info Ishaak wrote: Starting VPS ... vzquota : (error) Quota on syscall for 101: File exists vzquota on failed [3] Try stopping the VPS and dropping the quota file, then restarting both: vzctl stop 101 vzquota drop 101 vzquota on 101 vzctl start 101 -- Gregor Mosheh / Greg Allensworth System Administrator, HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] strange problem with nagios nrpe server
The good news is that I use Nagios with our VPSs, and it works brilliantly. include_dir=/etc/nagios/nrpe.d I have found that while this directive works under Xen this does not work under openvz. I find that surprising. Are you sure that the permissions didn't get mangled when you copied it over to ovz? That's the first thing I'd check: making sure that /etc/nagios/nrpe.d is in fact a directory, and that's readable by the user who runs nrpe (user nagios?). -- Gregor Mosheh / Greg Allensworth System Administrator, HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] I need info
I want implement server virtualization but i need to know the hardware requirements to install OpenVZ. Hard Disc space, RAM memory, processor speed, etc. so all Operating Systems installed run properly. I don't think that there are hard requirements. I successfully tested OpenVZ with several VPSs on a system with 1 GB of RAM and a 160 GB hard drive. The short version: a) If your hardware is compatible with the kernel (think Fedora or RedHat) then it's compatible with the OpenVZ kernel. The performance is not surprising at all: DDR2 RAM is faster than DDR, a Athlon X2 4000+ is faster than a Pentium-4, RAID rocks, and so on. b) Just plan for capacity, and remember that too much is always better than not enough. An extra GB of RAM, enough disk for the host OS plus some to spare, and you'll be A-OK. -- Gregor Mosheh / Greg Allensworth System Administrator, HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] Howto create a desired distro template?
Hi, guys. I've posted to the Wiki an abridged version of the document which I emailed last week, about creating a template cache for Slackware or HostGIS Linux. Some of it is a bit specialized, but hopefully it'll save someone some time. http://wiki.openvz.org/Creating_a_template_cache_:_Slackware_or_HostGIS_Linux#Zipping_it_up_into_a_cache_image Enjoy! And thanks again for OpenVZ. Our business couldn't do what we do, as well as we do, with VMWare or Xen. -- Gregor Mosheh / Greg Allensworth System Administrator, HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] Howto create a desired distro template?
Kir Kolyshkin wrote: Hi Gregor, Why don't you put it on wiki? Oh, I most assuredly will, as soon as my schedule permits. :) It'll be a few days as I get my new hardware together (so I can run 64-bit VMWare guests again) then a day or two as I apply the process and refine it. I'll likely get to it in the coming week. -- Gregor Mosheh / Greg Allensworth System Administrator, HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] Howto create a desired distro template?
DULMANDAKH Sukhbaatar wrote: It's nice document, is it possible to publish this on the wiki? and could you? I'll keep this email in my inbox, and will get to it when my schedule permits (a few days for my new hardware, a day or two as I tweak the procedure as I re-run it again for HostGIS Linux 4.2). I'll be glad to do it, but it'll be a week or so until I do. If you want it done sooner, you're welcome to start on it yourself. -- Gregor Mosheh / Greg Allensworth System Administrator, HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] Memory Error in VE starting TurboGears Project
guen wrote: Iam trying to runt a Python TurboGear Web-Project (including the TurboGear application server) in a VE, but I just get the error message "MemoryError". MemoryError - That's a clue. So are these fail counts: privvmpages 66 tcpsndbuf 9236 tcprcvbuf 2428 Up those limits and you should be A-OK. That limit is about 260 MB, and I notice that the limit in your conf file is about 200. If I recall, TurboGears can be a bit of a memory hog, and you'd be best increasing the limit until things run smoothly, before trying to lower the limits. -- Gregor Mosheh / Greg Allensworth System Administrator, HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] Howto create a desired distro template?
Hi, Donatas. I was faced with this same ordeal, as I am a OS creator/maintainer myself. Attached is our procedure, though you'll find that large parts of it don't apply to you (e.e. removing inetd and using xinetd instead, installing Nagios). Still, with this documentation and VMWare you should be on a good start as far as trimming down a VMWare machine running a OS, into a template cache. Cheers. -- Gregor Mosheh / Greg Allensworth System Administrator, HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA Creating a new "Host Template Cache" for HostGIS Linux This document describes how to use VMWare to create a new VM, install HostGIS Linux (HGL) on it and tweak the system into shape, and then create a Host Template Cache (a compressed VE image) for use in OpenVZ. * CREATING THE VM Start by creating a new VM in VMWare. The stats can be minimal, and there is no need to create the entire disk at once during the setup. * Create the disk as SCSI. Then install HGL. * Create a small partition at the end of the disk for swap. Some swap is technically necessary, but since you'll never in fact be using it, a few MB should be fine. * Do set the timezone properly. The internal clock does not use UTC/GMT. * Select the default mouse, but do NOT enable GPM at startup. * Hostname: template Domain: internal.lan * IP config: as appropriate for your LAN * Nameserver: no Reboot into your new HGL install. Now we want to tweak it into a usable template. Go ahead and login to the VM. * REPLACE INETD WITH XINETD removepkg inetd rm -f /etc/inetd.conf* /etc/rc.d/rc.inetd cd /tmp wget --header="Host: xinetd.org" http://204.152.188.37/xinetd-2.3.14.tar.gz tar zxvf xinetd*.gz cd xinetd* ./configure --prefix=/usr --sysconfdir=/etc make && make install mkdir /etc/xinetd.d cat >> /etc/rc.d/rc.local < /etc/xinetd.conf <> /etc/proftpd.conf echo "# keep all users chrooted to their homedir" >> /etc/proftpd.conf echo "DefaultRoot ~" >> /etc/proftpd.conf # allow the mailq to be checked by anybody: chgrp smmsp /var/spool/mqueue chmod g+rx /var/spool/mqueue # disable the root and user accounts # by changing the password for root and user to a ! character. vi /etc/shadow * UPGRADES AND SECURITY PATCHES The default HGL you used may require some software to be reinstalled, since new versions and critical bugfixes may have been released since that version of HGL was released. Follow these instructions, and also update them aas necessary for the appropriate versions and to remove paragraphs when a revision of HGL comes out that no longer requires them. /etc/rc.d/rc.pgsql stop cd /tmp wget --passive-ftp ftp://ftp.us.postgresql.org/pub/mirrors/postgresql/source/v8.2.4/postgresql-8.2.4.tar.bz2 tar jxvf postgresql-8.2.4.tar.bz2 cd postgresql-8.2.4 LDFLAGS=-lstdc++ ./configure --prefix=/usr --sysconfdir=/etc --localstatedir=/var \ --with-perl --with-python --with-openssl \ --enable-thread-safety --enable-integer-datetimes make && make install vi /var/lib/pgsql/postmaster.conf # set stats_row_collector=on * NAGIOS: THE HEALTH-MONITORING SYSTEM groupadd nagios useradd -g nagios -d /usr/local/nagios -m nagios echo "nrpe 5666/tcp # Nagios NRPE" >> /etc/services cd /tmp wget http://superb-east.dl.sourceforge.net/sourceforge/nagiosplug/nagios-plugins-1.4.6.tar.gz tar zxvf nagios-plugins-*.tar.gz ; cd nagios-plugins-* ./configure && make all && make install cd /tmp wget http://umn.dl.sourceforge.net/sourceforge/nagios/nrpe-2.6.tar.gz tar zxvf nrpe-2.6.tar.gz ; cd nrpe-2.6 ./configure && make && cp src/nrpe /usr/local/nagios/nrpe for plugin in \ check_wave check_users check_ups check_time check_tcp check_swap check_ssh check_ssmtp \ check_spop check_simap check_smtp check_sensors check_rpc check_real check_pop check_ping \ check_overcr check_oracle check_nwstat check_nt check_nntps check_nntp check_nagios \ check_mysql_query check_mrtgtraf check_mrtg check_log check_jabber check_ircd \ check_imap check_ifstatus check_ifoperstatus check_icmp check_http check_ftp check_flexlm \ check_file_age check_dummy check_disk_smb check_dig check_dhcp check_clamd check_by_ssh \ check_breeze check_apt check_udp do rm -f /usr/local/nagios/libexec/$plugin ; done cat > /usr/local/nagios/nrpe.cfg < /etc/xinetd.d/nrpe <> /etc/fstab echo "devpts /dev/pts devpts mode=0620 0 0" >> /etc/fstab # the startup sequence and services, even the firewall cd /etc/rc.d rm -f rc.gpm-sample rc.hotplug rc.ip_forward rc.modules rc.scanluns rc.serial rc.udev rc.sysvinit rc.firewall vi rc.syslog# delete all mentions of klogd vi rc.local
Re: [Users] reset user_beancounters?
Actually, I am curious about why resetting the UBCs is "an incorrect thing to do" The reason given is that apps may be reading them and may not be programmed to handle the number suddenly decreasing. So... What programs are these that are reading the UBCs? I haven't written any. and if I did I think it'd be great if the stats could be zeroed. Is the wiki page referring to some actually existing software which reads these stats which would be broken by a reset? I am confused as to why it would be "incorrect" to zero the stats, and why such concern is made for programs which don't exist or which we're not using. -- Gregor Mosheh / Greg Allensworth System Administrator, HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] forcibly purging quota?
Kirill Korotaev wrote: Can you please explain a bit how your quotas get wrong? Do you have some test case to reproduce the problem and demonstrate it? Well, now that you ask... I do not believe this to be a bug in vzquota. Many of our VPSs were P2V conversions, and I suspect that that's how the quotas started off badly. I SFTP'd the fles from their server on the HN, but didn't flush the quotas at the time. But, I am curious whether 'drop -f' and 'on -f' have any potential for mischief if used on a running VE. It would be nice to get the quotas updated without booting our customers. -- Gregor Mosheh / Greg Allensworth System Administrator, HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
[Users] forcibly purging quota?
Hey, guys. Thanks again for OpenVZ, and for the expert assistance. :) I have 15 VEs,and the quota is totally wrong on all of them. I have used the trick of "vzctl stop;vzquota drop;vzctl start" in order to restart a VE and have the quota properly updated. But... I would like to avoid shutting down our customers' VPSs. Question: How safe is "vzquota drop -f"? Could I use it to forcibly and safely have the quota refreshed, without shutting down each VE? What risks (if any) are involved? vzquota drop 1 -f vzquota on 1 -f -- Gregor Mosheh / Greg Allensworth System Administrator, HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] Physical server to OpenVZ
Stefan Kok wrote: 1) Is this possible (backup physical server and restore to virtual) ? 2) If so is there any documentation / HOWTO's or pointers that you could give please ? The basic OS would survive a tar/untar process. But there're some items typically missing from a VPS setting which would be vital to running on a physical server. The first that come to mind are module utilities (modprobe et al) and the modules themselves, some daemons (klogd, udevd), and then a few fundamental config files (fstab, lilo.conf or /boot). I think that you probably *could* do it with a lot of effort (90% of it being adding new packages to the VPS after you're untarred it), but that if your goal is fast and reliable recovery, you'd be best off restoring the VPS into a VPS setting. It's not difficult or time-consuming to install the OpenVZ system, and they could easily have it installed by the time the backup drive arrived at the new office. If the new HN already had OpenVZ, then it's trivially simple to run vzctl create to generate the VPS, then replace the content under /vz/private/X with the tar content. -- Gregor Mosheh / Greg Allensworth System Administrator, HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] error using vi in VE but no beans?
Since we're on the topic, I thought I'd pipe up for Steve's benefit about a small matter which I found baffling for a few days... If you run "df -k" in a VE, you will often get a different answer than vzquota. This is because vzquota doesn't keep to-the-moment statistics in its cache file. You must run "vzquota stat VEID" in order to have the cache updated; then vzquota will report the correct usage. This is in the Wiki, but I thought to save you some time. :) -- Gregor Mosheh / Greg Allensworth System Administrator, HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] error using vi in VE but no beans?
It turned out to be a disk quota issue. Sounds right. Glad you figured it out. BAD: the violation of the disk quota doesn't show up in user_beancounters Ah, quite true. At my work site, we use Nagios to monitor all of our VEs for general health, as we would any other systems: check the free space, make sure the database server is running, make sure that DNS is resolving, etc. Our VEs are leased out to untrusted customers who sometimes attempt amateur system administration, so some sanity/health checks are helpful on general principles. :) -- Gregor Mosheh / Greg Allensworth System Administrator, HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] error using vi in VE but no beans?
Steve Hodges wrote: E297: Write error in swap file E303: Unable to open swap file for "rc.local", recovery impossible "rc.local" 17 lines, 387 characters That's vi saying that it can't save a backup of the file (yeah, "swap file" really intuitive) That's probably just a permissions issue. Are you able to manually create and then delete a fake swap file named .rc.local.swp ? -- Gregor Mosheh / Greg Allensworth System Administrator, HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] openvz naming conventions; numeric vs symbolic
OK, I filed a bug for vzquota with minor severity, it will eventually be fixed: http://bugzilla.openvz.org/show_bug.cgi?id=668 Thanks again, Kir. OpenVZ is about the greatest thing we've ever found, and we're SO glad for it. -- Gregor Mosheh / Greg Allensworth System Administrator, HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] openvz naming conventions; numeric vs symbolic
Steve Wray wrote: Steve Wray wrote: There seems to be a slight inconsistency across the tool set here. vzctl does respect the given 'name' however vzquota does not appear to and seems to require the numeric id. Quite true. Did you check the bugtracker for the project, or log that as a bug? I'd love to see that fixed! -- Gregor Mosheh / Greg Allensworth System Administrator, HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] /dev/fd/3
Indeed. It sounds like a problem within the program. Something's causing a file descriptor to not be opened, and the program isn't performing sufficient error-checking. Cuz I know one can open file descriptors at will! To elaborate a bit more... The program "strace" is invaluable for finding out what a program is doing (or failing to do). If you strace your program's execution (and the failure is happening early enough) then you can probably find out which fopen() is failing and why. Like I said, this may be due to a VE issue (too many FDs open already, out of memory while reading file, etc) or may be something about the program itself. -- Gregor Mosheh / Greg Allensworth System Administrator, HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] /dev/fd/3
A program in a VE is complaining that it can't open /dev/fd/3 From reading, it appears that this is the equivalent of asking to open something with a file descriptor of 3. So I guess that the devices option on vzctl isn't going to help me here. Indeed. It sounds like a problem within the program. Something's causing a file descriptor to not be opened, and the program isn't performing sufficient error-checking. Cuz I know one can open file descriptors at will! Or perhaps your VE has maxed out the number of FDs it's allowed to have? Check the /proc/user_beancounters and see if anything there is happening to shed light on whether the failure is VE-related. I also find that whilst I have made /dev/hdc available to my VE, I can't mount the cdrom. It tells me mount: unknown filesystem type 'udf' mount: unknown filesystem type 'iso9660' For a CDROM, iso9660 is usually appropriate. Are you sure that that module is already loaded in the hardware node? A lot of distros leave iso9660 as a module and VEs cannot load kernel modules. Do a "modprobe iso9660" in the HN, to ensure that the module is loaded. Then try mounting the CD in the VE. -- Gregor Mosheh / Greg Allensworth System Administrator, HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] SATA HDD Problem
Markus Hardiyanto wrote: i check on this.. http://bugzilla.kernel.org/show_bug.cgi?id=8650 it seems that it the same problem that i encounter.. how to solve this? Ouch, that's a rough one! The bugzilla shows that they're actively working on it as of yesterday, and it sounds as if they're some ways away from a solution. Once they figure it out, it may be some time before the fix makes it into the official Linux kernel, and longer still until it gets into a OpenVZ kernel. If you're okay with applying the OpenVZ patches to your own kernel source and building your own OpenVZ-capable kernel, your best bet may be to sit tight and keep watching that bugzilla page until they fix this bug in the kernel. Then grab that newest source and give it a try... Sorry I don't have anything more constructive to say, but that seems to be the status at the moment. :( -- Gregor Mosheh / Greg Allensworth System Administrator, HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] Creating more permanent veth`s
Macka Radek wrote: I have followed how&to in your wiki http://wiki.openvz.org/Virtual_Ethernet_device but with no sucess. It seems there is problem in script /usr/sbin/vznetaddoute in line VZHOSTIF=`echo $NETIF |sed 's/^.*host_ifname=\(.*\),.*$/\1/g'` I am not able to resolve what author of the script wants use as VZHOSTIF. Hi, Macka. You're in luck; I wrote that new script, and it works for me. This is under FC6, not Debian, but that shouldn't make a difference. The vznetaddroute script checks the VE's conf file for the NETIF line, which should be very similar to this except for the MACS: NETIF="ifname=eth0,mac=00:18:51:D2:F4:A5,host_ifname=veth12.0,host_mac=00:18:51:1F:70:28" The VZHOSTIF is the device name given as host_ifname in that string, see? In this case, veth12.0 Personally, I've found it a very handy practice to always keep the ifname the same as the VEID, e.g. veth1.0 for VE 1, veth72.0 for VE 72, and so on. But that's up to you. If your VE's conf is missing tyhat NETIF line line, you probably didn't add a veth to it in the first place, so do this: vzctl set $VEID --save --netif_add eth0 And then don't forget to add the two new lines to the config, manually: CONFIG_CUSTOMIZED="yes" VETH_IP_ADDRESS="12.34.56.78" Let us know if you need more help! -- Gregor Mosheh / Greg Allensworth System Administrator, HostGIS cartographic development & hosting services http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] Re: strange issue with mysql crashing in openvz
Catam: Where did you dig that up, and do you know why it's the case? Is this something about OpenVZ in particular, or Xeons in particular, ... ? catam wrote: > Seems I have found the issue , was CFLAGS > /proc/cpuinfo => > processor : 0 > vendor_id : GenuineIntel > cpu family : 15 > model : 4 > model name : Intel(R) Xeon(TM) CPU 3.00GHz > you need > CFLAGS="-march=nocona -O2 -pipe -fomit-frame-pointer" > I had > CFLAGS="-march=k8 -O2 -pipe" ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] VEs and rebooting
>> >> I've noticed that there's a /etc/cron.d/vz, which runs >> >> /usr/share/vzctl/scripts/vpsreboot >> Sounds good but... what exactly is the point of this functionality? Vasily Tarasov wrote: > For example the root of VE performed `reboot` command inside VE. VE will > stop, but somebody from outer world should start it again. vpsreboot is > this somebody, who starts VE if it was rebooted. ;) Ahhh, okay. Running "init 6" in a VE stops the VE's processes and shuts it down, but has no mechanism of restarting it. Then a few minutes later vpsreboot boots it back up, simulating the rebooting part of a reboot. Of course! Thanks for the explanation; it makes sense now that I have the missing piece filled in. :) I'll make a note of this for the template caches I'm creating. ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] VEs and rebooting
> Ian P. Christian wrote: >> I've noticed that there's a /etc/cron.d/vz, which runs >> /usr/share/vzctl/scripts/vpsreboot >> My understanding is that vpsreboot should scan for >> /vz/private/nnn/reboot, and if it finds that file, it will start the VPS >> and delete it the file. Pardon me if this question sounds dense. I'm looking at the vpsreboot code, since I'd like to understand everything there is to know about VZ The vpsreboot cronjob runs every 5 minutes. It looks for a /reboot file in each VE's directory; if it exists and the VE is stopped, it starts the VE. Sounds good but... what exactly is the point of this functionality? Is it used as part of the VZ startup sequence, or for watchdogging, ...? -- HostGIS Cartographic development and hosting services 707-822-9355 http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] new version of vznetaddroute script
> Gregor Mosheh wrote: >> Hey, guys. I've come up with a much more flexible version of the >> vznetaddroute script described in the Wiki Kir Kolyshkin wrote: > You don't have to ask anybody -- just register on wiki.openvz.org, login > and make the changes! Yeah, I knew that I *could* just do it, but wasn't sure of the policy in that regard, and I'd hate to seem disrespectful. But if it's open policy, then I'll go for it later today, once I get my morning email cleared up. -- HostGIS Cartographic development and hosting services 707-822-9355 http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] error from RkHunter and ChkRootKit
> Markus Hardiyanto <[EMAIL PROTECTED]> writes: >> I install RkHunter and ChkRootKit inside VE. >> Performing 'known good' check... >> /bin/kill [ BAD ] (etc) > Yes and no -- those are modified from the standard packages you would > have in a normal system, but the modification is to be expected with Thanks for the reply on this. Starting over the next few days, I was about to implement our policy of using chkrootkit and rkhunter on customer VEs. so this was good to know ahead of time. What is the nature of the modifications, and to which files? We're running Slackware (actuallly, HostGIS Linux, which is Slackware-based) using a hand-made template cache I made per some directions I found in the Wiki. So if some binaries need to be modified, that'd be good to know too. >> from ChkRootKit: >> Checking `lkm'... You have74 process hidden for readdir command >> chkproc: Warning: Possible LKM Trojan installed > Again, probably expected: the proc file system within the VE isn't > identical to a physical system. Right. Still a scary message to receive. What is the nature of the discrepancy here? What would show up in the VE's /proc area that wouldn't also show up in their ps output? More importantly: Is it even possible for a VE to load a kernel module at all? Or is a LKM check completely irrelelvant in a VE context? -- HostGIS Cartographic development and hosting services 707-822-9355 http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
[Users] new version of vznetaddroute script
Hey, guys. I've come up with a much more flexible version of the vznetaddroute script described in the Wiki http://wiki.openvz.org/Virtual_Ethernet_device#Making_a_veth-device_persistent Improvements in my script are: - multiple IPs can be specified! hello, ip aliases in VEs! - nicer syntax and error checking And actually, I find that entire section to be not particularly easy to read and follow. Who do I contact for permission to rewrite that section? -- HostGIS Cartographic development and hosting services 707-822-9355 http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] trouble booting with FC6-64
Kir Kolyshkin wrote: > Yum doesn't know anything special about the kernels ... > So it's up to the user to decide which kernel to use. ... > # yum install ovzkernel[-flavor] Hey, I bet that's it. I bet he didn't specify "ovzkernel-smp" Thanks for the tip, Kir. -- HostGIS Cartographic development and hosting services 707-822-9355 http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] setting memory allocation
Andrew Cranson (Layershift Limited) wrote: > I believe you can also use M and G > suffixes on recent OVZ package releases Good to hear someone else saying this; I wondered if I was just crazy for thinking M and G, that I'm imagined seeing them somewhere. ;) On my very-outdated test system for my initial experiments, those didn't work at all. But I'd love to see it work using a modern version of OpenVZ on modern hardware. :) -gregor ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] trouble booting with FC6-64
The latest: During his experiments last night with CentOS, he installed the kernel and it actually frooze during boot. He then manually downloaded and installed the SMP kernel, and it worked A-OK. So, we're wondering whether "'yum install ovzkernel" was installing an inappropriate kernel for a 8-CPU machine. Kirill Korotaev wrote: > You may need to press Alt-SysRq-P and Alt-SysRq-T and collect the output. > this will dump the required info about whether kernel hangs. I'll let him know. Since I always work on remote systems, I've never even gotten to use the magic key! Vasily Tarasov wrote: > what OpenVZ kernel version do you use? > Have you compiled the kernel by your self or you have used a rpm > package? > What FC6 kernel version does boot smoothly? I wrote up the install procedure for 32-bit systems, but now he's trying it on a 64-bit system, so... The troublesome one is the RPM, installed via "yum install ovzkernel" under FC6-64 Doing this same thing on 32-bit works perfectly. ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] setting memory allocation
I'd like to take a shot at answering this, to "quiz myself" on how well I understand this stuff. So if my answers are incorrect or incomplete, please speak up! > I have a computer with P4 2,4Ghz processor and 1GB of RAM. I'm planning to > split it into 3 VEs with this memory allocation: > VE1: 512MB > VE2: 256MB > VE3: 256MB If by "memory allocation" you mean "the amount of RAM they're guaranteed to have available for use by apps" then try this: # RAM is 4k pages, so 131072*4k = 512M vzctl set 1 --save --vmguarpages 131072 vzctl set 1 --save --oomguarpages 131072 vzctl set 1 --save --privvmpages 131072:196608 But, if you really have only 1 GB of RAM, it may not be wise to allocate all of it. If all of the VEs really use all their RAM, the system will start swapping to make up more RAM (e.g. for the HN's own use) and nobody enjoys a system that's swapping. -Gregor ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
[Users] trouble booting with FC6-64
Hi, all. I wrote a few days ago, to the effect that the fellow setting up our hardware was having trouble getting OpenVZ to work under FC6 on x86_64. Using the same procedure, he can use FC6 32-bit without a hitch. He's trying to boot into the OpenVZ kernel, and it's stopping. Iit gets as far as "Input: AT Translated set 2 keyboard as /class/input/input0" which is the keyboard loading. That should be followed immediately by the root filesystem... but it's not. Just sits there. He says that it doesn't panic, and that it's not frozen cuz a ctrl-alt-del will reset the machine. Just gets to that point and stops. Any thoughts on what could cause that or how we can try to get to the bottom of it? If necessary, I can get a hardware list. Again, it works A-OK under FC6 32-bit, which is what makes this so mysterious. -- HostGIS Cartographic development and hosting services 707-822-9355 http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
[Users] trouble booting with FC6-64
Hi, all. I wrote a few days back that the fellow setting up our hardware is having problems -- HostGIS Cartographic development and hosting services 707-822-9355 http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] Re: FW: 64-bit OpenVZ
> Hey, guys. > Here's the tech's concern: >> Here is the link that describes a work around for the openvz problem. Sorry - I forgot to mention this probably-relevant fact: I have no intention of using vzrpm or vzyum. I prefer to work within the VE, so I'd either use "vzctl exec" or just enter the VE and run yum as usual. ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
[Users] Re: FW: 64-bit OpenVZ
Hey, guys. Here's the tech's concern: > Here is the link that describes a work around for the openvz problem. Note > that fedora core 6 64bit has not been tested. The first paragraph titled > "THE PROBLEM" described what I said: > > http://wiki.openvz.org/Install_OpenVZ_on_a_x86_64_system_Centos-Fedora Any comments? ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] Re: 64-bit stability?
Benny Amorsen wrote: > The OpenVZ kernel parts work very nicely on 64-bit. We have both > 32-bit and 64-bit guests on a 64-bit kernel. And Kirill said the same. This is what I had suspected, I was just surprised to hear the tech report "a problem" I'll have him get more specific, and may end up hand-walking him through the installation via email. *smirk* Thanks a lot, guys. -- HostGIS Cartographic development and hosting services 707-822-9355 http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
[Users] 64-bit stability?
Hi, all. How stable is OpenVZ considered on 64-bit platforms, specifically Intel Xeons? I did some searches on the site, but couldn't find anything solid on the subject, only mention of "fixes" I ask because the tech who's setting up our hardware "had a problem with installing OpenVZ" If OpenVZ on 64-bit is known to be solid, then I'll prod him for some more specific errors and see whether we can get this to happen. -- HostGIS Cartographic development and hosting services 707-822-9355 http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] PostgreSQL can't allocate shared memory
> right. Can you please describe this on wiki? Sure thing; will do this evening. Thanks for the clarifications; and thanks for the correction that the HN's shmmax is really for the HN and not for the VEs. ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] PostgreSQL can't allocate shared memory
>> And the HN's kernel.shmmax is set to 1 GB. (Is that relevant in a VE?) > check the same in VE please. > AFAICS, by default kernel sets: > #define SHMMAX 0x200 > which is 32Mb. > So you have to increase it in VE /etc/sysctl.conf file or in /proc. Ahhh, very tricky -- in that it's exactly what one should expect! Thanks, Kirill; I can't believe that this escaped me. I had thought that the VE would have shmmax set by the UBC; evidently it's the case that each VE gets their own set of sysctls for perfectly ordinary use. So, let me summarize the shared memory situation for VEs: * The HN's kernel.shmmax sets an absolute limit on shared memory, for both the HN and any VEs. * A VE's shmpages sets the limit on shared memory for that VE. Therefore it must be <= kernel.shmmax * Then the VE uses sysctl kernel.shmmax to set their allowable limit, which of course must be <= the VE's shmpages. Sound right? -- HostGIS Cartographic development and hosting services 707-822-9355 http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
[Users] PostgreSQL can't allocate shared memory
Hi, all. First: OpenVZ; nice. A slick alternative to VMWare, less overhead. Nice. But, a hitch in my first experiments... I have PostgreSQL 8.1 in a VE (Fedora Core 6, for what it's worth), and I cannot set shared_buffers higher than 3700 (about 30 MB) and have it start. Anything higher and it complains that it can't allocate the shared memory. In the UBC, I verified that I set shmpages to 32000, which should be 128 MB at 4k apiece (or 256 MB if it's 8k pages; I'm not sure which it is for shared memory pages) and therefore more than enough for PgSQL's request. I have tried raising shmpages, but it doesn't change anything. The privvmpages is 262144 (1 GB) so this should also be plenty. The UBC is showing 0 failcount across the board. And the HN's kernel.shmmax is set to 1 GB. (Is that relevant in a VE?) I'd be very appreciative of any help. And thanks for the software! -- HostGIS Cartographic development and hosting services 707-822-9355 http://www.HostGIS.com/ "Remember that no one cares if you can back up, only if you can restore." - AMANDA ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
