[OpenAFS] openafs-server does not recover from crash

Mark Vitale mvitale@sinenomine.net
Wed, 9 Mar 2022 18:09:35 +0000


Pascal,

> On 9 Mar 2022, at 11:26 AM, Pascal Salet <pascal.salet@wu.ac.at> wrote:
>=20
> our openafs-server has stopped working after a crash.
>=20
> "bos status" shows all services online for all fileservers and DBservers.
>=20
> udebug port 7003 works correctly from all fileservers and DBservers.
>=20
> However, "strace -p $(pidof salvageserver)" shows an error:
> connect(6, {sa_family=3DAF_UNIX, sun_path=3D"/var/lib/openafs/local/fssyn=
c.sock"}, 110) =3D -1 ECONNREFUSED (Connection refused)
>=20
> SalsrvLog:
> Wed Mar 09 16:09:31 2022 @(#)OpenAFS 1.8.2-1-debian 2018-09-12

Unfortunately this version of OpenAFS has the Rx CID bug:=20
  http://openafs.org/frameset/dl/openafs/1.8.7/RELNOTES-1.8.7

This bug may only become apparent on the first restart after Jan 15, 2021.
You must upgrade all your clients and servers to OpenAFS 1.8.7 or higher.

Regards,
--
Mark Vitale
mvitale@sinenomine.net