Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
Am Montag, 29. August 2005 18:22 schrieb Jeff Fisher:
> Hey all,
>
> I know there is already a thread on this; however, I just joined the list.
>
> I had been using drbd 0.7.11 on two machines (P4 3Ghz machines w/ 1 gig
> of ram each and a SCSI hardware RAID1 hdd setup -- of which 50gig is
> setup for drbd) and I was use to seeing 38,000 to 42,000 K/sec sync
> speeds on the initial sync.
>
> I upgraded to drbd 0.7.12 and to test I invalidated the remote side and
> started a full sync. The fastest sync speed I saw was 88 K/sec.
>
> I am using the same config with both 0.7.11 and 0.7.12 and the syncer
> option is set to 50M.
>
> Any ideas?
The thing is, that I can not reproduce the problem here.
I have to say that the permanent test cluster has rather old hardware,
I usually get here a resync throughput of 9MB/sec (100 MBit network...)
and this works just as usual with 0.7.12 for me.
We have now Jeff and Andre, with the symptom that the syncer is slow
or even comes to a halt.
1) Please find out if it is just slow, or if it halts.
E.g. by looking at the numbers in the parentheses in the line
with the syncer progress bar.
2) In case it comes to a halt (I think it halts) please
do "echo t > /proc/sysrq-trigger" on both maschines.
This will print a stack trace of all processes to the
kernel log.
This is really a lot of text. Please post only the stack
traces of the relevant processes (those with a drbd in
their name) to the list.
Thanks!
-Philipp
--
: Dipl-Ing Philipp Reisner Tel +43-1-8178292-50 :
: LINBIT Information Technologies GmbH Fax +43-1-8178292-82 :
: Schönbrunnerstr 244, 1120 Vienna, Austria http://www.linbit.com :