[DRBD-user] DRBD Crashing/Stalling on Sync

Tom Pawlowski tpawlowski at fortressitx.com
Tue Sep 29 23:18:22 CEST 2009

Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.


Version: 8.0.14 (api:86)
GIT-hash: bb447522fc9a87d0069b7e14f0234911ebdab0f7 build by
phil at fat-tyre, 2008-11-12 16:40:33

That's what comes with Debian 5.0.3. Really best not to use it, since
it's missing certain configuration directives that later versions use
that can help with tweaking performance (rcvbuf-size but not
sndbuf-size, etc.)

Regards,
--T

On Tue, Sep 29, 2009 at 5:01 PM, Adam Taylor <adam.taylor at wml.co.nz> wrote:
> :) Good stuff.  I have seen the same issue on a few installs were I have had
> to set sync rate to about the same which seems to resolve the initial sync
> issue.  What version were you using?
>
> Cheers
>
> Adam
>
> -----Original Message-----
> From: tpawlowski.itx at gmail.com [mailto:tpawlowski.itx at gmail.com] On Behalf
> Of Tom Pawlowski
> Sent: Wednesday, September 30, 2009 8:00 AM
> To: Adam Taylor
> Cc: drbd-user at lists.linbit.com
> Subject: Re: [DRBD-user] DRBD Crashing/Stalling on Sync
>
> Hey Adam,
>
> The VPS gets hammered pretty hard, so I've had to throttle the sync rate
> down pretty far. I believe when this happened, it was around 15-20MB/s on a
> Gigebit crossover cable.
>
> That being said, I managed to successfully run a sync last night without any
> interruption or stalls at the same speeds. So, that's a relief. :)
>
> More than likely, I guess the only solution would have been to update DRBD
> if it persisted. I think from here on out I'll deploy newer versions than
> what comes with Debian Lenny.
>
> Thanks folks!
>
> --T
>
> On Tue, Sep 29, 2009 at 4:50 PM, Adam Taylor <adam.taylor at wml.co.nz> wrote:
>> Hi Tom,
>>
>> Have you tried lowering your Sync Rate in drbd.conf?
>>
>> Thanks
>>
>> Adam
>>
>> -----Original Message-----
>> From: drbd-user-bounces at lists.linbit.com
>> [mailto:drbd-user-bounces at lists.linbit.com] On Behalf Of Tom Pawlowski
>> Sent: Tuesday, September 29, 2009 8:41 AM
>> To: drbd-user at lists.linbit.com
>> Subject: [DRBD-user] DRBD Crashing/Stalling on Sync
>>
>> Hi all,
>>
>> I've come across an issue on a clustered set up that I haven't been
>> able to find a resolution for. (It's a bit different than the other
>> stalled ticket as neither of the nodes are locking up on their own.
>> They remain stable and are able to ping each other on both eth0/eth1.)
>> If anyone could shed some light on it, I'd very much appreciate it.
>>
>>
>> *SETUP*
>>
>> Two hardware nodes, running a DRBD/LVM/Xen stack (in that order). One
>> node is using a RAID-1 3ware controller with two Western Digital Blue
>> 1.0TB drives, the other RAID-5 3ware controller with three WD Black RE
>> 500GB drives. Everything else is identical.
>>
>> (I'm attempting to sync the data to the RAID-5 node so I can failover
>> to that and replace RAID-1 on the other node--given that the
>> performance with DRBD on the latter is awful.)
>>
>> Controller information:
>>
>> node1
>> 3w-9xxx: scsi0: Found a 3ware 9000 Storage Controller at 0xda100000, IRQ:
>> 16.
>> 3w-9xxx: scsi0: Firmware FE9X 4.06.00.004, BIOS BE9X 4.05.00.015, Ports:
> 4.
>>
>> node2
>> 3w-9xxx: scsi0: Found a 3ware 9000 Storage Controller at 0xda100000, IRQ:
>> 16.
>> 3w-9xxx: scsi0: Firmware FE9X 4.06.00.004, BIOS BE9X 4.05.00.015, Ports:
> 2.
>>
>> Software Version information for both nodes are identical:
>>
>> Distro: Debian 5.0.3
>> Kernel: Linux 2.6.26-2-xen-amd64 #1 SMP Thu Aug 20 00:36:34 UTC 2009
>> x86_64 GNU/Linux
>> drbdadm Version: Version: 8.0.14 (api:86) Xen Version: 3.0.3
>>
>>
>> *PROBLEM*
>>
>> The sync will run along without any issues for a few hours, after
>> which /proc/drbd reports that it has stalled. The Xen instance running
>> on top stops responding entirely. There are a number of stack traces
>> in the system log which I have attached to the email.
>>
>> Has anyone come across something like this before? Updating DRBD is a
>> bit iffy, as our client is very downtime-adverse given that he's
>> paying for a high-availability setup. Don't know if that can be done
>> without reinitializing both resources.
>>
>> Thanks in advance for any help!
>>
>> Regards,
>> Tom Pawlowski
>>
>>
>>
>> __________ NOD32 4468 (20090929) Information __________
>>
>> This message was checked by NOD32 antivirus system.
>> http://www.eset.com
>>
>>
>
> __________ NOD32 4468 (20090929) Information __________
>
> This message was checked by NOD32 antivirus system.
> http://www.eset.com
>
>
>



More information about the drbd-user mailing list