XenServer 5.6 FP1 DRBD working

Xenlogo

Hi all,

The wait is over so to speak, XenServer 5.6 FP1 was released over a month ago but DRBD wouldn’t work due to a bug in blktap2,

The issue was highlighted by Tomoe Sugihara on the Xen-Api mailing list prior to FP1 being released – http://lists.xensource.com/archives/html/xen-api/2010-12/msg00001.html

There is mention of a couple of patches but neither of them fixed the issue – it was necessary to download the latest Soruce from the Xen Cloud Project release 1 RC1 and use the blktap2 source files.

The steps to make it work, Very simple,

1. Download the patched Kernel RPM (Link below)

2. Use WinSCP or similar to load it onto your XenServer

3. Login as root on the console

4. Run # rpm -ivh kernel-2.6.32.120.7.1.xs5.6.100.307.170586xen-1.i386.rpm –force

5. Reboot

This will replace the current kernel with the patched one.

That’s it DRBD is working now

Enjoy and let me know how it goes.

Cheers, Joe

If you already have DRBD installed and setup before installing the Kernel – you will need to reinstall the DRBD RPM, modprobe & start DRBD

Just run

# rpm -ivh drbd-km-2.6.32.12_0.7.1.xs5.6.100.307.170586xen-8.3.10-1.i386.rpm –force

# /etc/init.d/drbd start

# modprobe drbd

Direct link download Kernel + Utilities: http://tinyurl.com/6xr4g6h

For installing DRBD with XenServer including the latest release 8.3.10 please see http://joekane.eu/xenserver-drbd-update-4-the-guide

41 thoughts on “XenServer 5.6 FP1 DRBD working

  1. I’ve just confirmed that DRBD 8.3.8.1 and FP1 is working with the supplied patch. Live migration and all the other features are working as expected. Much thanks!!

  2. So it looks like I jumped the gun, upgrading XenServer wiped out DRBD and its configuration files. XenServer booted up, mounted the disks and ran like nothing had changed. There is no sign of the drbd kernel module running. It wasn’t until a moment after the cluster diverged that disk errors cropped up during more testing (they weren’t really in sync). You will need a recompiled DRBD rpm for FP1.Moral of the story is, don’t get so excited you overlook the obvious things even if it is a test environment, lol.

  3. Yes if you install the new kernel after installing DRBD you need to reinstall the DRBD rpm – I have found the drbd.conf is still ok and the metadata is intact.Added some notes to the post.

  4. Hi I have xenserver 5.6 FP1 installation working with drbd and live migration is working but I am experiencing some packet lost when a migrate a virtual machine to a different physical host. I run "ping virtualmachineDomain" and the virtual machine keeps replying ok. Then I perform the live migration and I lose around 5-10 icmp replies.┬┐people using drbd is experiencing this behaviour? ┬┐or should I have no packet loss? Not sure if this can be related to the config in switches ┬┐maybe something about the arp table not refreshing ok? Both physical nodes are connected to the same switch in different ports and both ports are in trunk mode. This is the config in those ports:interface GigabitEthernet0/8 switchport mode trunk┬┐anyone knows any other thing I could check/try? many thanks in advance for any help.

  5. I went to do this upgrade, but I cannot find the download for drbd-utils-8.3.10 can you repostThanks,Sam Kuonen

  6. No worries though, I just downloaded drbd83-utils-8.3.10-1.el5.elrepo.i386.rpm and installed the kernel module with the –nodeps flag.BTW, Thanks for all your hard work!

  7. @aid, download the patched kernel and kernel module, then install drbd83-utils-8.3.10-1.el5.elrepo.i386.rpm from ELrepo.orgYou can install the kernel module (drbd-km) package with –nodeps optionIt all worked for me. I’m already running a fresh install in production.

  8. Hi Aid,Just tried both links and all is working ok – the files are hosted by 2shared & uploading.com so you can expect some advertising i guessI have added a direct link for both the Kernel & Utilitieshttp://tinyurl.com/5rm4fel

  9. Joe I managed to download DRBD and utils (ZIP) ok from your latest link. Could you do the same for the kernel? Thanks.

  10. Hello Joseph,do I understand it right? That *.zip containts the patched kernel plus the compiled drbd kernel module?Or should I have to install drbd like this:- download your patches kernel and install it- install drbd by hacking yum –nopgpcheck — nodeps –enablerepo=extras install drbd83 kmod-drbd in the console?Please help my to solve this issueKind regardsJohn

  11. Hi John,Yes the ZIp file contains the Patched Kernel, DRBD & DRBD utilitiesWinSCP the 3 files onto your XenServer e.g. /root/kernel-2.6.32.120.7.1.xs5.6.100.307.170586xen-1.i386.rpm/root/drbd-km-2.6.32.12_0.7.1.xs5.6.100.307.170586xen-8.3.10-1.i386.rpm/root/drbd-utils-8.3.10-1.i386.rpmcd /rootrpm -ivh kernel-2.6.32.120.7.1.xs5.6.100.307.170586xen-1.i386.rpm –forceRebootThen install DRBD using this guide http://joekane.eu/xenserver-drbd-update-4-the-guideHope this helps.Cheers, Joe

  12. Ok, got it.But I don’t have to install DRBD using your guide (http://joekane.eu/xenserver-drbd-update-4-the-guide), is that correct? Is it possible to install all the *.zip’s files and DRBD would work ootb?Like this (after unpacking the zip and uploading it via WinSCP):cd /rootrpm -ivh kernel-2.6.32.120.7.1.xs5.6.100.307.170586xen-1.i386.rpm –forcerebootrpm -ivh drbd-km-2.6.32.12_0.7.1.xs5.6.100.307.170586xen-8.3.10-1.i386.rpm –forcerpm -ivh kernel-2.6.32.120.7.1.xs5.6.100.307.170586xen-1.i386.rpm –forcerpm -ivh drbd-utils-8.3.10-1.i386.rpm –forcecp /usr/share/doc/drbd-utils/drbd.conf /etc/drbd.confdoing some customizationsmodprobe drbd/etc/init.d/drbd startfinishThanks in advance and cheersJohn

  13. One more thing, Joseph:Does your made drbd-utils package contain heartbeat? If not what are the steps to implement this nice feature for XEN Server 5.61 FP1?Great blog, I love to read your stuff :)Cheers, John

  14. Hi Joe,I just set up my first test of DRBD with Xenserver 5.6 FP1. I got as far a running modprobe and I get "FATAL: Module drbd not found." I used your Zip file containing the Patched Kernel, DRBD & DRBD utilities. Any ideas on troubleshooting this before I wipe everything and start over?Thanks

  15. Hi Paul,Try installing DRBD again using,mv /etc/drbd.conf /etc/drbd.conf.backuprpm -ivh drbd-utils-8.3.10-1.i386.rpm –forcerpm -ivh drbd-km-2.6.32.12_0.7.1.xs5.6.100.307.170586xen-8.3.10-1.i386.rpm –forcerm -f /etc/drbd.confmv /etc/drbd.conf.backup /etc/drbd.confmodprobe drbdLet me know how you get on.Cheers, Joe

  16. Hi John,Late reply sorry about that – I haven’t go around to implementing heartbeat+XenServerIts certainly possible but I suspect it would require a couple of custom scripts.Cheers, Joe

  17. That did the trick, I’m syncing one of my 1800GB virtual disks now.Due to Xenserver not working well with virtual disks over 1800 GB I have two virtual disks at/under that size (sdb and sdc) that I would like to use as one SR. How would I implement this with DRBD? Also, what is the correct value for the syncer with 10G cards (identical 10G’s in each server connected by crossover cable)? I tried 10G, but it appears the maximum value is 4G.Thanks for you help!

  18. Hi Joe,after several attempts i still got this error:-( /dev/sda3 is bigger then 1800GB, should that be the Problem?!xe sr-create shared=true device-config:device="/dev/drbd1" name-label="DRBD-SR1" type=lvmError code: SR_BACKEND_FAILURE_55Error parameters: , Logical Volume partition creation error [opterr=error is 255]XenServer 5.6 FP1, DRBD 8.3.10 ( http://joekane.eu/xenserver-56-fp1-drbd-working )already done:fdisk /dev/sdad – Delete the partition /sda3w – write the changesFdisk /dev/sdan – add a new partition Choose defaultsw – write changesdd if=/dev/zero of=/dev/sda3 bs=4096 count=1000Server-Infos:/dev/disk/by-idtotal 0lrwxrwxrwx 1 root root 9 Apr 20 17:16 edd-int13_dev80 -> ../../sdalrwxrwxrwx 1 root root 10 Apr 20 17:16 edd-int13_dev80-part1 -> ../../sda1lrwxrwxrwx 1 root root 10 Apr 20 17:16 edd-int13_dev80-part2 -> ../../sda2lrwxrwxrwx 1 root root 10 Apr 20 17:16 edd-int13_dev80-part3 -> ../../sda3lrwxrwxrwx 1 root root 9 Apr 20 17:16 scsi-36001517faa9600001536d9870a6b4d32 -> ../../sdalrwxrwxrwx 1 root root 10 Apr 20 17:16 scsi-36001517faa9600001536d9870a6b4d32-part1 -> ../../sda1lrwxrwxrwx 1 root root 10 Apr 20 17:16 scsi-36001517faa9600001536d9870a6b4d32-part2 -> ../../sda2lrwxrwxrwx 1 root root 10 Apr 20 17:16 scsi-36001517faa9600001536d9870a6b4d32-part3 -> ../../sda3lrwxrwxrwx 1 root root 10 Apr 20 17:16 usb-Optiarc_DVD_RW_AD-7700S_000000000033 -> ../../scd0xe sr-list params=alluuid ( RO) : 339b7126-d9c1-aedb-4521-e9ec064e8d8c name-label ( RW): DVD drivesuuid ( RO) : cd61c22c-b942-093d-4f2e-97738af2e5e4 name-label ( RW): Removable storage host ( RO): guardian allowed-operations (SRO): forget; VDI.introduce; plug; update; destroy; scan; VDI.clone; unpluguuid ( RO) : 0b455e9f-9ee5-9ed9-55a1-ccf5040d626c name-label ( RW): Removable storage host ( RO): colossus fdisk -lDisk /dev/sda: 1995.9 GB, 1995997970432 bytes Device Boot Start End Blocks Id System/dev/sda1 * 1 523 4194304 83 LinuxPartition 1 does not end on cylinder boundary./dev/sda2 523 1045 4194304 83 Linux/dev/sda3 1045 242666 1940826005+ 83 Linux cat /proc/drbdversion: 8.3.10 (api:88/proto:86-96)GIT-hash: 5c0b0469666682443d4785d90a2c603378f9017b build by root@localhost.localdomain, 2011-02-08 17:00:08 1: cs:Connected ro:Primary/Primary ds:UpToDate/UpToDate C r—– ns:307200 nr:0 dw:204800 dr:102400 al:25 bm:7 lo:0 pe:0 ua:0 ap:0 ep:1 wo:b oos:0common { protocol C; startup { become-primary-on both; } disk { max-bio-bvecs 1; } net { allow-two-primaries; cram-hmac-alg "sha1"; shared-secret "foo123"; after-sb-0pri discard-zero-changes; after-sb-1pri consensus; after-sb-2pri disconnect; } syncer { rate 1G; } handlers { # split-brain "/usr/lib/drbd/notify-split-brain.sh root"; }}resource drbd-sr1 { on colossus { device /dev/drbd1; disk /dev/sda3; address 10.1.1.1:7788; meta-disk internal; } on guardian { device /dev/drbd1; disk /dev/sda3; address 10.1.1.2:7788; meta-disk internal; }}

  19. It is my understanding that Xenserver itself does not work with virtual disks over 1800GB. You may be able to create two virtual disks of that size or smaller and use them together in one SR, but your virtual disk or disk partition size should not exceed 1800GB.

  20. /dev/sda3 1045 209670 1675785635+ 83 Linuxshould be small enough…but i still get: xe sr-create shared=true device-config:device="/dev/drbd1" name-label="drbd-sr1" type=lvmError code: SR_BACKEND_FAILURE_55Error parameters: , Logical Volume partition creation error [opterr=error is 255],

  21. Hello,i have a DRBD Pool with 2 Nodes running since ~90 Days with a Backup script which takes snapshots of the VM┬┤s and export it to a share on a NAS.I was wondering of the size of the DRBD Storage. The VM┬┤s take about ~130 GB of space on the DRBD Storage but XenCenter says that there is ~250 GB of ~350 GB allocated. On the CLI i got many "unused" or "NOT AVAILIBLE" LV┬┤s when i perform a lvdisplay like this. — Logical volume — LV Name /dev/VG_XenStorage-50beb20c-2d46-f2af-28ef-5d97a59fea92 /VHD-65befd9b-c5b2-4074-aa1b-77077afac80f VG Name VG_XenStorage-50beb20c-2d46-f2af-28ef-5d97a59fea92 LV UUID c7Dseg-2qfg-XTXa-zkhv-w3Dv-hxTg-29iE1I LV Write Access read/write LV Status NOT available LV Size 38.08 GB Current LE 9749 Segments 2 Allocation inherit Read ahead sectors autoI hope that this Hotfix (http://support.citrix.com/article/CTX129267) will apply with my problem but i dont know if the DRBD will work after i install the hotfix.Is there anyone with the same issuses or knows howto fix it?Thanks for the answers.Chris

Leave a Reply

Your email address will not be published. Required fields are marked *