[ClusterLabs] Antw: Re: EL6, cman, rrp, unicast and iptables

Noel Kuntze noel at familie-kuntze.de
Mon Sep 14 11:41:02 UTC 2015


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

Hello Ullrich,


> Then that's not FIFO, but priority scheduling. Eveybody knows the starvation problem of priotity scheduling.
It's mixed. The individual bands behave like a FIFO. The bands are prioritized over each other.

> Imagine some cluster filesystem has it's own timeount / fencing mechanism. Then TOTEM when going wild can cause starvation of other services. It's the wrong design IMHO.
> I wonder whether this can explain the mysterous cLVM retransmit list growing under some loads.
>
> [...]
It can, if the total bytes in the queue is higher than the transmission rate of the network interface.
Packets in band 1 and 2 are only not delivered if band 0 always has packets. But again, for that, the transmission rate must be lower than
the rate at which the queues grows.
Does totem generate a runaway stream of packets when they are delivered in time?

 

- -- 

Mit freundlichen Grüßen/Kind Regards,
Noel Kuntze

GPG Key ID: 0x63EC6658
Fingerprint: 23CA BB60 2146 05E7 7278 6592 3839 298F 63EC 6658

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2
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=qEtz
-----END PGP SIGNATURE-----






More information about the Users mailing list