----- Forwarded message from Andrej Hosna <hosna@example.com> -----
Date: Tue, 26 Sep 2000 22:27:22 +0200
From: Andrej Hosna <hosna@example.com>
To: Fabio Olive Leite <olive@example.com>
Subject: Re: DRBD problems
X-Mailer: KMail [version 1.0.29.2]
Hello Fabio,
> This is strange, which version of drbd are you using?
I run 0.57
> Is the device mounted on the primary when you do this?
Yes, it's mounted on primary and it's NOT mounted on secondary
> What does /proc/drbd say? Are the devices really connected?
If I check right before i try to shut down secondary node, the connection is
there, and both statuses are all right.
> I think we'll need some more information about your setup and about
> what exactly you do when it hangs.
I have not experienced this only with one cluster but my current configuration
is:
Aramis: AMD K6-3 , 128 MB RAM,
Adino: Celeron 400, 128 MB RAM,
100 Mbit Ethernet connection , rate on drbd set to 3000
cca 100MB partition dedicated for drbd
IDE disks, with no RAID.
> ) Seems like there is some infinite loop somehwere within
> ) kernel and it between CLI and STI.
>
> Are there any kernel messages? Any kernel panic?
No kerenel messages, no kernel panic ...
kernel is just frozen ... So there must be problem somewhere in interupt driver
or something like this ...
> ) Also I experienced problems with running drbd with MOSIX kernels , mut maybe
> ) it's the same thing.
> Well, I've never run drbd on a Mosix kernel, but it would be nice to
> hear your experiences.
After some trials I have definitively get rid of MOSIX.... You need to set up
it very precisely to get it work fine ... Otherwise it can happen that part of
your heartbeat deamon processes will run on secondary node , and probably U
don't want this. Anyhow , MOSIX can be very useful , IF configured as I said
VERY carefuly.
I hope that this clean up situation a bit ....
Go ahead and contact me, I'm eager to find out what 's the cause of this.
p.s. this problem doesn't appear everytime, it's just occasional.
Regards,
Andrej
--
/* Andrej Hosna - http://people.ibl.sk/adino - +421 903 852 696 */
/* IBL Software Engineering - http://www.ibl.sk - +421 7 43427214 */
----- End forwarded message -----
--
( Fábio Olivé Leite -- Conectiva HA Team -- olive@example.com )
( PPGC/UFRGS MSc candidate -- Advisor: Taisy Silva Weber )
( Linux - Distributed Systems - Fault Tolerance - Security - Pizza )
Date: Tue, 26 Sep 2000 22:27:22 +0200
From: Andrej Hosna <hosna@example.com>
To: Fabio Olive Leite <olive@example.com>
Subject: Re: DRBD problems
X-Mailer: KMail [version 1.0.29.2]
Hello Fabio,
> This is strange, which version of drbd are you using?
I run 0.57
> Is the device mounted on the primary when you do this?
Yes, it's mounted on primary and it's NOT mounted on secondary
> What does /proc/drbd say? Are the devices really connected?
If I check right before i try to shut down secondary node, the connection is
there, and both statuses are all right.
> I think we'll need some more information about your setup and about
> what exactly you do when it hangs.
I have not experienced this only with one cluster but my current configuration
is:
Aramis: AMD K6-3 , 128 MB RAM,
Adino: Celeron 400, 128 MB RAM,
100 Mbit Ethernet connection , rate on drbd set to 3000
cca 100MB partition dedicated for drbd
IDE disks, with no RAID.
> ) Seems like there is some infinite loop somehwere within
> ) kernel and it between CLI and STI.
>
> Are there any kernel messages? Any kernel panic?
No kerenel messages, no kernel panic ...
kernel is just frozen ... So there must be problem somewhere in interupt driver
or something like this ...
> ) Also I experienced problems with running drbd with MOSIX kernels , mut maybe
> ) it's the same thing.
> Well, I've never run drbd on a Mosix kernel, but it would be nice to
> hear your experiences.
After some trials I have definitively get rid of MOSIX.... You need to set up
it very precisely to get it work fine ... Otherwise it can happen that part of
your heartbeat deamon processes will run on secondary node , and probably U
don't want this. Anyhow , MOSIX can be very useful , IF configured as I said
VERY carefuly.
I hope that this clean up situation a bit ....
Go ahead and contact me, I'm eager to find out what 's the cause of this.
p.s. this problem doesn't appear everytime, it's just occasional.
Regards,
Andrej
--
/* Andrej Hosna - http://people.ibl.sk/adino - +421 903 852 696 */
/* IBL Software Engineering - http://www.ibl.sk - +421 7 43427214 */
----- End forwarded message -----
--
( Fábio Olivé Leite -- Conectiva HA Team -- olive@example.com )
( PPGC/UFRGS MSc candidate -- Advisor: Taisy Silva Weber )
( Linux - Distributed Systems - Fault Tolerance - Security - Pizza )