Hello, I tried to research about something like lost data on DomU, didn`t find any clue. My electricity power went out for server and I started server couple hours after and I got lost on DomU 2 month data. On Dom0 seems to be all ok, on DomU was data before last reboot, two month ago. Dom0 and DomU is on the same discs, on raid1(mirror) only separated by partitions. DomU on top lvm and ext3 fs used in xen with phy discs. Don`t use snapshotting, Debian 6 os with xen 4.0.1 server installed from original packages. Raid discs all working ok, no other clue what went wrong. Two month black hole. Seems it was not hacked because it lost data around loosing electricity time and packages in this working time was updated, but now it has two month old again and need updates again. Any idea what went wrong with fs and how it could lost data till last DomU reboot(not Dom0)? Configuration is simple and used many years on many servers... Thanks, Casper
Niels Dettenbach (Syndicat IT&Internet)
2012-Aug-23 19:27 UTC
Re: DomU loses 2 month data after lost electricity
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 Casper <kl@os.lv> schrieb:>Any idea what went wrong with fs and how it could lost data till last >DomU reboot(not Dom0)? >Configuration is simple and used many years on many servers...Thus hardly depend from your application in detail, but it should very less depend from xen what kind of data you will lost with a OS crash like yours. Crashing a DomU is mainly like crashing a real OS. Depending from your apps it may be that filesys buffers over longer time are not written / synced to disc. You may try to seek in your swap against some helpful data (even this depends from your app) and if you have no snapshots or backups in between there is nothing other to save anymore. best regards, Niels. - -- Niels Dettenbach Syndicat IT&Internet http://www.syndicat.com -----BEGIN PGP SIGNATURE----- Version: APG v1.0.8 iIEEAREIAEEFAlA2hB06HE5pZWxzIERldHRlbmJhY2ggKFN5bmRpY2F0IElUJklu dGVybmV0KSA8bmRAc3luZGljYXQuY29tPgAKCRBU3ERlZRyiDTs9AJ9vwpkOXXUh p/GYLOsNCUYWRbepfwCfTP4bK5Mo9/9TX+9UeG2XUyejS9M=AYyc -----END PGP SIGNATURE-----
It`s web development server, postgresql, svn, all usual apps, I think not hard used development. How can maybe I detect problem with filesystem buffers? All is usual sw raid mirror, lvm partitions on top ext3 with DomU...> > On 08/23/2012 10:27 PM, Niels Dettenbach (Syndicat IT&Internet) wrote: >> -----BEGIN PGP SIGNED MESSAGE----- >> Hash: SHA256 >> >> >> >> Casper <kl@os.lv> schrieb: >> >>> Any idea what went wrong with fs and how it could lost data till last >>> DomU reboot(not Dom0)? >>> Configuration is simple and used many years on many servers... >> >> Thus hardly depend from your application in detail, but it should very >> less depend from xen what kind of data you will lost with a OS crash >> like yours. >> >> Crashing a DomU is mainly like crashing a real OS. >> >> Depending from your apps it may be that filesys buffers over longer >> time are not written / synced to disc. >> >> You may try to seek in your swap against some helpful data (even this >> depends from your app) and if you have no snapshots or backups in >> between there is nothing other to save anymore. >> >> >> best regards, >> >> Niels. >> - -- >> Niels Dettenbach >> Syndicat IT&Internet >> http://www.syndicat.com >> -----BEGIN PGP SIGNATURE----- >> Version: APG v1.0.8 >> >> iIEEAREIAEEFAlA2hB06HE5pZWxzIERldHRlbmJhY2ggKFN5bmRpY2F0IElUJklu >> dGVybmV0KSA8bmRAc3luZGljYXQuY29tPgAKCRBU3ERlZRyiDTs9AJ9vwpkOXXUh >> p/GYLOsNCUYWRbepfwCfTP4bK5Mo9/9TX+9UeG2XUyejS9M>> =AYyc >> -----END PGP SIGNATURE----- >> >>