Re: [DRBD-user] DRBD Sync stalls
On Mon, Aug 03, 2009 at 09:25:09PM -0700, Димитър Бойн wrote: Hi, again! So hoping that the new DRBD 8.3 might address my issue I did the upgrade. :-( it didn't help. As I did not receive any answers to my original post 2 weeks ago I guess my problem is unique? There have been two answers. What would be the best way to troubleshoot? Any way to troubleshoot? have a look at my sig ;) -- : Lars Ellenberg : LINBIT HA-Solutions GmbH : DRBD®/HA support and consultinghttp://www.linbit.com DRBD® and LINBIT® are registered trademarks of LINBIT, Austria. __ please don't Cc me, but send to list -- I'm subscribed ___ drbd-user mailing list drbd-user@lists.linbit.com http://lists.linbit.com/mailman/listinfo/drbd-user
Re: [DRBD-user] slow drbd over tripple gigabit bonding balance-rr
Mark, syncer rate does not define the overall drbd performance between the two nodes, but rather it designates a limit for re-syncing, to assure normal system performance while it is happening in the backgrounnd (so your re-sync processes dont eat up all your bandwidth between your drbd nodes, and normal drbd usage can happen). It has nothing to do with the working performance of a single drbd device pair see more here: http://www.drbd.org/users-guide/s-configure-syncer-rate.html Originally it was set to be 80M in this case, but probably it should be even lower, since despite the iperf results, I never saw a drbd go over the 117MB/s (single gigabit link hitting performance wall), so it probably should be as low as 35MB/s As I wrote before, the two boxes have three gigabit links dedicated for DRBD, these are bonded using the balance-rr mode, and with arp ip monitoring (basically I can unplug any of the cables between the nodes, in any order, and plug them back in, as long there is a single link, the connection is uninterrupted, also depending how many gigabit links are allive the bandwidth scales up with every additional connection, pretty swift actually) What I would like to see is higher write rates. I know how the different bondings work, I also know balance-rr is the only where a single connection cab scale beyond the capacity of a single card, and it clearly happens when I benchamrk it with iperf, but never saw it in DRBD. Now, since this is a Xen Dom0, I have been able to do more testing in the DomU itself (half of the testing was done before I wrote to the mailing list) In the paravirtualized DomU, the drbd devices are inported as xvdb to xvdf, and they were used: 1) as phisical volumes for the LVM in the DomU itseld, setting up the logical volumes using striping in LVM 2) as part of a raid0 stripe 3) raid0 stripe in above as a phisical volume for LVM There is no signofocant difference in either mode. I have even tested a RAID_0 stripe over the drbd volumes in the Xen Dom0, same results. I also know about the LVM default block device performance issues, and blockdev --setra is used as a workaround on all levels Whenever I disconnect drbd, performance is as expected (close to the raid10 performance measured) The reason this is so annoying, is because on both the DRBD wiki and in the mail list there are hints that even over a dual gigabit link in ballance-rr performance is much better, see them yourself: http://www.drbd.org/home/wiki/?tx_drwiki_pi1[keyword]=performance http://www.nabble.com/DRBD-Performance-td18745802.html http://lists.linbit.com/pipermail/drbd-user/2008-July/009893.html Also if it not clear, this is a nested LVM (with the DRBD's being the phisical volumes for LVM in the paravirtualized xen instance, DRBD is running in Dom0): DRBD-| |-xvdb (PV)| Xen Dom0 6XHDD-RAID10-LVM-DRBD-|XenDomU|-xvdc (PV)|-LVMfile ystems DRBD-| |-xvdd|(PV)| As a side note, I am a seasoned sysadmin of fifteen years and use linux practically for everything for the last ten years, work with it at least twelve hours a day (mostly more than that, I am lucky to do for living what I love, and work and fun are the same) So, anybody knows how those magical numbers were achived under those links? z On Thu, Jul 30, 2009 at 9:18 AM, Mark Watts m.wa...@eris.qinetiq.comwrote: On Thu, 2009-07-30 at 03:57 -0400, Zoltan Patay wrote: using dd if=/dev/zero of=/dev/drbd26 bs=10M count=100 I get: drbd connected 1048576000 bytes (1.0 GB) copied, 13.6526 seconds, 76.8 MB/s 1048576000 bytes (1.0 GB) copied, 13.4238 seconds, 78.1 MB/s 1048576000 bytes (1.0 GB) copied, 13.2448 seconds, 79.2 MB/s drbd disconnected 1048576000 bytes (1.0 GB) copied, 4.04754 seconds, 259 MB/s 1048576000 bytes (1.0 GB) copied, 4.06758 seconds, 258 MB/s 1048576000 bytes (1.0 GB) copied, 4.06758 seconds, 258 MB/s The three (intel) gigabit PCIe cards are bonded with balance-rr, and iperf gives me: iperf 0.0-10.0 sec 2.52 GBytes 2.16 Gbits/sec (276.48MB/s) So clearly there is enough speed for both on the network and in the backend to support higher speeds. The boxes are with cross-over back-to-back no-switch. version: 8.3.0 (api:88/proto:86-89) GIT-hash: 9ba8b93e24d842f0dd3fb1f9b90e8348ddb95829 build by p...@fat-tyre, 2008-12-18 15:26:13 global { usage-count yes; } common { syncer { rate 650M; } } Try actually setting this to a sensible value for 3 x 1Gbit links. eg: 300M resource OpenVZ_C1C2_B_LVM5 { protocol C; startup {degr-wfc-timeout 120;} disk {on-io-error detach;no-disk-flushes;no-md-flushes;no-disk-drain;no-disk-barrier;} net { cram-hmac-alg sha1; shared-secret OpenVZ_C1C2_B; allow-two-primaries; after-sb-0pri discard-zero-changes; after-sb-1pri discard-secondary; after-sb-2pri disconnect; rr-conflict disconnect;
[DRBD-user] SOLVED - kind of (Fwd: slow drbd over tripple gigabit bonding balance-rr)
turns out, once I did the testing with 1M block size instead of 10M, it was showing the performance I expected. sync; echo 3 /proc/sys/vm/drop_caches # free pagecache, dentries and inodes sync;dd if=/dev/zero of=/vz/blob bs=1M count=1000 1000+0 records in 1000+0 records out 1048576000 bytes (1.0 GB) copied, 4.04935 seconds, 259 MB/s I wonder if max-buffers and max-epoch-size set to 2048 has anything to do with this. z PS: I want to thank Mark for replying to the post. -- Forwarded message -- From: Zoltan Patay zoltanpa...@gmail.com Date: Thu, Jul 30, 2009 at 3:57 AM Subject: slow drbd over tripple gigabit bonding balance-rr To: drbd-user@lists.linbit.com using dd if=/dev/zero of=/dev/drbd26 bs=10M count=100 I get: drbd connected 1048576000 bytes (1.0 GB) copied, 13.6526 seconds, 76.8 MB/s 1048576000 bytes (1.0 GB) copied, 13.4238 seconds, 78.1 MB/s 1048576000 bytes (1.0 GB) copied, 13.2448 seconds, 79.2 MB/s drbd disconnected 1048576000 bytes (1.0 GB) copied, 4.04754 seconds, 259 MB/s 1048576000 bytes (1.0 GB) copied, 4.06758 seconds, 258 MB/s 1048576000 bytes (1.0 GB) copied, 4.06758 seconds, 258 MB/s The three (intel) gigabit PCIe cards are bonded with balance-rr, and iperf gives me: iperf 0.0-10.0 sec 2.52 GBytes 2.16 Gbits/sec (276.48MB/s) So clearly there is enough speed for both on the network and in the backend to support higher speeds. The boxes are with cross-over back-to-back no-switch. version: 8.3.0 (api:88/proto:86-89) GIT-hash: 9ba8b93e24d842f0dd3fb1f9b90e8348ddb95829 build by p...@fat-tyre, 2008-12-18 15:26:13 global { usage-count yes; } common { syncer { rate 650M; } } resource OpenVZ_C1C2_B_LVM5 { protocol C; startup {degr-wfc-timeout 120;} disk {on-io-error detach;no-disk-flushes;no-md-flushes;no-disk-drain;no-disk-barrier;} net { cram-hmac-alg sha1; shared-secret OpenVZ_C1C2_B; allow-two-primaries; after-sb-0pri discard-zero-changes; after-sb-1pri discard-secondary; after-sb-2pri disconnect; rr-conflict disconnect; timeout 300; connect-int 10; ping-int 10; max-buffers 2048; max-epoch-size 2048; } syncer {rate 650M;al-extents 257;verify-alg crc32c;} on c1 { device /dev/drbd26; disk /dev/mapper/xenvg-OpenVZ_C1C2_B_LVM5; address10.0.10.10:7826; meta-disk /dev/mapper/xenvg-DRBD_MetaDisk[26]; } on c2 { device/dev/drbd26; disk /dev/mapper/xenvg-OpenVZ_C1C2_B_LVM5; address 10.0.10.20:7826; meta-disk /dev/mapper/xenvg-DRBD_MetaDisk[26]; } } Some of the settings above are unsafe (no-disk-flushes;no-md-flushes), they were turned on to see if it makes any different (did not) The two boxes are quad core 3GHz Nehalems, 12GB tripple channel DDR3-1600, 6 western digital caviar black 750GB hdds, in RAID10 with LVM on top of it, the DRBD backends are carved out of LVM. Three separate Intel gigabit PCIe cards are bonded with ballance-rr, and connects them back-to-back, with a forth gigabit card in each box (onboard) toward the outside. The OS is Debian Etch + Backports with some custom deb packages rolled by me. The machine is a Xen Dom0, kernel: 2.6.26, xen: 3.2.1, drbd: 8.3.0 Thanks any help / hint in advance, z ___ drbd-user mailing list drbd-user@lists.linbit.com http://lists.linbit.com/mailman/listinfo/drbd-user
[DRBD-user] I think that there is a missing instruction for RPM building.
Hi all. I tried to follow the directions in the URL: http://www.drbd.org/users-guide/s-build-rpm.html and it seems that something is missing: Before executing make rpm there is a need to build the .filelist. This can be done by make .filelist or by make tarball. Thanks Dotan ___ drbd-user mailing list drbd-user@lists.linbit.com http://lists.linbit.com/mailman/listinfo/drbd-user
[DRBD-user] switch over takes a quite long time
Hello, context : 3-node cluster, every node connected, HA services on node1, DRBD version 8.3.2 on linux 2.6.30. I switch HA services over to node2 with /usr/lib/heartbeat/hb_standby all from node1. It takes a long time to perform : 20 seconds. I have this in logs (ha-log) : on node1 : ResourceManager[7475]: 2009/08/04_11:53:27 info: Running /etc/ha.d/resource.d/drbdupper r0-U start Filesystem[7799]: 2009/08/04_11:53:47 INFO: Resource is stopped ResourceManager[7475]: 2009/08/04_11:53:47 info: Running /etc/ha.d/resource.d/Filesystem /dev/drbd1 /data ext3 start and on node2 : heartbeat[4836]: 2009/08/04_11:53:26 info: Local standby process completed [all]. heartbeat[4836]: 2009/08/04_11:53:26 info: New standby state: 3 heartbeat[4836]: 2009/08/04_11:53:26 info: Managed go_standby process 12561 exited with return code 0. heartbeat[4836]: 2009/08/04_11:53:48 WARN: 1 lost packet(s) for [node1] [7702:7704] heartbeat[4836]: 2009/08/04_11:53:48 info: remote resource transition completed. Question : why drbdupper takes such a long time to start? Is that normal? Thanks. here is the drbd.conf file : global { usage-count yes; } common { syncer { rate 10M; } net { max-buffers 4; } } resource r0 { protocol C; handlers { pri-on-incon-degr echo o /proc/sysrq-trigger ; halt -f; pri-lost-after-sb echo o /proc/sysrq-trigger ; halt -f; local-io-error echo o /proc/sysrq-trigger ; halt -f; } startup { wfc-timeout 0; degr-wfc-timeout 120; } disk { on-io-error detach; } net { after-sb-0pri disconnect; after-sb-1pri disconnect; after-sb-2pri disconnect; rr-conflict disconnect; } syncer { rate 90M; al-extents 128; csums-alg md5; } on node1 { device /dev/drbd0; disk /dev/md2; address10.0.0.1:7788; meta-disk /dev/md1 [0]; } on node2 { device /dev/drbd0; disk /dev/md2; address10.0.0.2:7788; meta-disk /dev/md1 [0]; } } resource r0-U { protocol C; syncer { csums-alg md5; rate 5M; } stacked-on-top-of r0 { device/dev/drbd1; address 192.168.2.15:7788; } on node3 { device/dev/drbd1; disk /dev/md2; address 192.168.2.14:7788; meta-disk internal; } } ___ drbd-user mailing list drbd-user@lists.linbit.com http://lists.linbit.com/mailman/listinfo/drbd-user
Re: [DRBD-user] Antwort: upgrade problems
Hi, Thanks for your replies. I followed your replies and it worked like you described. Unfortunately I could not get to finish the upgrade because the HDD died when upgrading the second server so I had to create everything from start. On Mon, Jul 27, 2009 at 1:34 PM, robert.koe...@knapp.com wrote: Hi, the message you get during upgrade is perfectly normal. answer with yes and it search for old metadata, answer with yes again if it offers to upgrade ant it will upgrade the metadata. however be sure to have a working backup of yout data in case something goes wrong. DO NOT ANSWER WITH YES IF NO V7 METADATA IS FOUND! I have upgraded about half a dozen DRBD clusters in the lost months following this guide: http://blogs.linbit.com/florian/2007/10/03/step-by-step-upgrade-from-drbd-07-to-drbd-8/ Mit freundlichen Grüßen / Best regards, Robert Köppl System Administration --- Phone: +43 3842 805-910 Fax: +43 3842 805-500 robert.koe...@knapp.com www.KNAPP.com --- KNAPP Systemintegration GmbH Waltenbachstrasse 9 8700 Leoben, Austria --- Commercial register number: FN 138870x Commercial register court: Leoben --- ___ drbd-user mailing list drbd-user@lists.linbit.com http://lists.linbit.com/mailman/listinfo/drbd-user ___ drbd-user mailing list drbd-user@lists.linbit.com http://lists.linbit.com/mailman/listinfo/drbd-user
Re: [DRBD-user] split brain detected when switching back to the 2node cluster from the DR node
Hi, There are several items such as below in /etc/drbd.conf. You may need to study it. My DRBD is 8.3.0. It works well. edward #after-sb-0pri disconnect; after-sb-0pri discard-older-primary; #after-sb-1pri disconnect; after-sb-1pri discard-secondary; Date: Tue, 4 Aug 2009 18:58:15 +0200 From: pierre.lebr...@laposte.net To: drbd-user@lists.linbit.com Subject: [DRBD-user] split brain detected when switching back to the 2node cluster from the DR node Hello, I always get a split brain when I switch the HA services back to the 2node cluster from my DR node. Here are the steps I follow : - HA services are on the DR node - I stop these HA services - I umount the data The state of DRBD on node3 is as follow : -- version: 8.3.2 (api:88/proto:86-90) GIT-hash: dd7985327f146f33b86d4bff5ca8c94234ce840e build by r...@hcns1, 2009-08-04 09:41:09 1: cs:WFConnection ro:Primary/Unknown ds:UpToDate/DUnknown C r ns:0 nr:34083396 dw:34084032 dr:68168163 al:12 bm:2094 lo:0 pe:0 ua:0 ap:0 ep:1 wo:f oos:208 -- Then, on node1 : - drbdadm primary r0 - I start the HA IP - drbdadm --stacked up r0-U At this point, every thing is OK. Here is the output of cat /proc/drbd : -- version: 8.3.2 (api:88/proto:86-90) GIT-hash: dd7985327f146f33b86d4bff5ca8c94234ce840e build by r...@hans1, 2009-08-04 09:43:39 0: cs:Connected ro:Primary/Secondary ds:UpToDate/UpToDate C r ns:13706 nr:174 dw:15096 dr:102401147 al:30 bm:2170 lo:0 pe:0 ua:0 ap:0 ep:1 wo:f oos:0 1: cs:Connected ro:Secondary/Primary ds:UpToDate/UpToDate C r ns:0 nr:244 dw:244 dr:416 al:0 bm:9 lo:0 pe:0 ua:0 ap:0 ep:1 wo:d oos:0 -- Then, I set all things back (reset) : on node1 : - drbdadm --stacked down r0-U - drbdadm secondary r0 - I stop the HA IP The state on node1 is as follow : -- version: 8.3.2 (api:88/proto:86-90) GIT-hash: dd7985327f146f33b86d4bff5ca8c94234ce840e build by r...@hans1, 2009-08-04 09:43:39 0: cs:Connected ro:Secondary/Secondary ds:UpToDate/UpToDate C r ns:13707 nr:174 dw:15097 dr:102401147 al:30 bm:2186 lo:0 pe:0 ua:0 ap:0 ep:1 wo:f oos:0 1: cs:Unconfigured -- on node3, I type these commands to reset the state : - drbdadm secondary r0-U Then, on node1 and node2, I start heartbeat normally. Well, each time I follow theses steps, node3 gets a split-brain. Where is the problem? context : 3-node cluster, every node connected, HA services on node1, DRBD version 8.3.2 on linux 2.6.30. ___ drbd-user mailing list drbd-user@lists.linbit.com http://lists.linbit.com/mailman/listinfo/drbd-user _ Share your memories online with anyone you want. http://www.microsoft.com/middleeast/windows/windowslive/products/photos-share.aspx?tab=1___ drbd-user mailing list drbd-user@lists.linbit.com http://lists.linbit.com/mailman/listinfo/drbd-user