[pve-devel] pve-kernel-4.4.10-1-pve woes
Alexandre DERUMIER
aderumier at odiso.com
Mon Jul 4 06:10:49 CEST 2016
Hi,
do you have any stack trace ?
Hopefully, I'm stil on 4.4.8 and no crash for me.
I have only 1 server running on 4.4.10, but without nfs (only rbd), and I don't have crash.
I have have seen any change for nfs in 4.4.9 and 4.4.10 changelog, so maybe it's NIC related
(they are 1 or 2 patch for infiniband)
https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.4.9
https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.4.10
----- Mail original -----
De: "datanom.net" <mir at datanom.net>
À: "pve-devel" <pve-devel at pve.proxmox.com>
Envoyé: Dimanche 3 Juillet 2016 20:58:40
Objet: [pve-devel] pve-kernel-4.4.10-1-pve woes
Hi all,
After facing problems with mysterious "crashes" of proxmox nodes I have
been doing some extensive tests this weekend. My tests have shown that
this kernel has introduced a regression concerning NFS. The problems is
must visible on Infiniband but Intel ethernet nics are also heart by
this regression.
How to reproduce:
Put a lot of load on a NFS mount like making backup, creating a CT or
creating a VM.
What will happen:
The nic(s) involved will stop passing traffic through and only way to
bring them back up is ifdown nic && ifup nic. Problem with this for NFS
mounts is that NFS in its default configuration will wait indefinitely
because of mount option 'hard' bringing the hole node down which can
only be resolved by a reboot of the node.
Because of this I have reverted to pve-kernel-4.4.8-1-pve which works
as expected.
--
Hilsen/Regards
Michael Rasmussen
Get my public GnuPG keys:
michael <at> rasmussen <dot> cc
http://pgp.mit.edu:11371/pks/lookup?op=get&search=0xD3C9A00E
mir <at> datanom <dot> net
http://pgp.mit.edu:11371/pks/lookup?op=get&search=0xE501F51C
mir <at> miras <dot> org
http://pgp.mit.edu:11371/pks/lookup?op=get&search=0xE3E80917
--------------------------------------------------------------
/usr/games/fortune -es says:
The trouble with being punctual is that nobody's there to appreciate it.
-- Franklin P. Jones
_______________________________________________
pve-devel mailing list
pve-devel at pve.proxmox.com
http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
More information about the pve-devel
mailing list