[DRBD-user] drbd0: could not kmalloc() nbh

Tim Tregubov tim at zingweb.com
Sat May 29 21:55:39 CEST 2004

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


I tried the tar again and this time I'm getting tons of:
EXT3-fs error (device drbd(43,0)): ext3_free_blocks: bit already 
cleared for block 8489630

HUH? Corruption?  Yikes!

searching the list for solutions... and can't find anything. There is a 
recent thread with the same error message as this one, but with 
different reasons I think.

This is just corruption right?  I tried running fsck and it found so 
many things wrong with duplicate blocks and stuff that it was freaky!

Any chance anybody is reading this over the weekend? :-)

Cheers,
Tim




On May 29, 2004, at 1:23 PM, Tim Tregubov wrote:

> Hiya -
>
> Sorry if this has already been asked (I searched the lists and didn't 
> find a resolution per say).
>
> Is seeing lots and lots of this bad:
> kernel:  drbd0: could not kmalloc() nbh
> last message repeated 713 times
> ?
>
> It was preceeded by a sync:
> drbd0: Synchronisation started blks=15
> drbd0: Synchronisation done.
>
>
> I'm running: version: 0.6.12 (api:64/proto:62)
> on:   2.4.21-15.ELsmp   (RH ES 3 Taroon Update 2)
> compiled with: gcc (GCC) 3.2.3 20030502 (Red Hat Linux 3.2.3-34)
> filesystem is ext3
>
> At the time this happened I was running a 9.7GB untar onto the file 
> system (shortly after the full sync).
>
> The messages only appeared on one of the two systems.
>
> Help!   I need to put this system into production monday morn, and I'm 
> stressing. :-)
>
> Thanks so much for any help!
>
> Cheers,
> Tim
>
>
>
>
> the drbd.conf:
> resource drbd0 {
>
>   protocol=C #used to be B but seems like C is prefered now
>   fsckcmd=/bin/true #used to be fsck -p -y but we're on ext3 now
>   inittimeout=120
>
>   disk {
>         do-panic #for io-errors added
>         disk-size=40025916
>   }
>
>   net {
>     tl-size     = 8000
>     sync-group  = 0
>     sync-min    = 1M
>     sync-max    = 600M # if you don't care about network saturation
>     timeout     = 120   # unit: 0.1 seconds
>     connect-int = 30    # unit: seconds
>     ping-int    = 30    # unit: seconds
>     ko-count    = 10     # if some block send times out this many 
> times,
>                         # the peer is considered dead, even if it still
>                         # answeres ping requests
>   }
>
>   on shadowfax {
>     device=/dev/nb0
>     disk=/dev/hdb1
>     address=10.12.13.1
>     port=7789
>   }
>
>   on gwaihir {
>     device=/dev/nb0
>     disk=/dev/hdc1
>     address=10.12.13.2
>     port=7789
>   }
> }
>
> _______________________________________________
> drbd-user mailing list
> drbd-user at lists.linbit.com
> http://lists.linbit.com/mailman/listinfo/drbd-user
>




More information about the drbd-user mailing list