Data Protector - Passion for Data Protection

data-protector.org

Home » HowTo » Hanging Sessions with DP 7.01

Hanging Sessions with DP 7.01

Posted on by 6 Comments

It has been told to me and I’ve seen it in some environments… with DP 7.01 it can occur that a sessions hang till forever or the backup is aborted after 7200 second. The message displayed shows IPC errors (i.e error below) as no more data from Disk Agent or Media Agent is sent to Cell Server in this case, in general this might be a network related problem. I’ve seen this error in environments with firewalls which disconnet TCP sessions when no data is transmitted, however, this can be normal from backup point of view. As a solution you can look at omnirc file, from my point of view following variables have to be enabled anyway, for any further recommendation please try DPcfgCreator by Stephan Voss.

Update: The variables OB2IPCKEEPALIVETIME and OB2IPCKEEPALIVEINTERVAL are valid on Windows only, for Linux and HP-UX the TCP behaviour is modified by changing kernel parameters and with the omnirc variable OB2IPCKEEPALIVE. For more information, please refer to http://h30499.www3.hp.com/t5/Data-Protector-Support-and-News/Data-Protector-Support-Tip-Linux-Only-Using-KeepAlive-to-resolve/td-p/6263015#.UuC5thQwc08. Thanks to Alex for clarifying that.

OB2SHMIPC=0
OB2IPCKEEPALIVE=1
OB2INETTIMEOUT=10
OB2IPCKEEPALIVETIME=900
OB2IPCKEEPALIVEINTERVAL=60

[Major] From: BSM@< client > "Exchange"  Time: 11.09.2012 01:00:55
[61:3003]             Lost connection to OB2BAR Backup DA named "ERROR"
                on host < client >.
                Ipc subsystem reports: "IPC Read Error
                System error: [10054] Verbindung wurde von Peer zurückgesetzt"


[Major] From: BSM@< client > "Exchange"  Time: 11.09.2012 01:00:55
[61:3003]             Lost connection to OB2BAR Backup DA named "ERROR"
                on host < client >.
                Ipc subsystem reports: "IPC Read Error
                System error: [10054] Connection reset by peer"

6 thoughts on “Hanging Sessions with DP 7.01”

  2. Hi Patrick,

    hast du einen Call dazu offen?
    Die Parameter haben bei uns leider nur temporär (3 Backups lang) Abhilfe geschaffen. Dann kamen die IPC Errors zurück. Warum ist mir schleierhaft.
    Es scheint nur BAR Backups (Exchange 2010 in unserem Fall) zu betreffen.

    Ich habe dazu mittlerweile einen Call offen, mal schauen was dabei herumkommt!

  3. Könnt ihr bitte schauen, ob bei euch auch der VBDA auf dem fraglichen Client hängt wenn der Job beendet wird? Habe das bisher nur bei BAR Backups gesehen.

    Grüße,
    Sebastian

  5. Hi,

    in meiner omnirc.tmpl (noch nie genutzt) ist keiner der Parameter bis jetzt enthalten.
    Allerdings haben wir auch noch die DP 6.0 am laufen.

    Kann/soll ich diese Parameter trotzdem nutzen? (und werden Sie mir überhaupt was bei den Fehlern s.u. bringen?)

    Wir haben seit einigen Tagen nämlich erhebliche Schwierigkeiten mit den Fehlermeldungen:

    [Critical] From: BMA-NET@DPServer.intern \”IBM:ULTRIUM-TD4_1_DPServer\” Time: 24.03.2013 07:28:56
    IPC Handle Table Full (Overflow)

    [Warning] From: BDA-NET@Fileserver.intern \”D:\” Time: 24.03.2013 07:12:28
    Connection to BMA on host DPServer.intern lost, trying to reconnect.

    Viele Grüße
    Steffen

  6. Dear all,

    variables OB2IPCKEEPALIVETIME and OB2IPCKEEPALIVEINTERVAL only work on Windows platforms, using Linux/unix please read the following article and change system default tcp behaviour for tcp_keepalive_time after setting omnirc variable OB2IPCKEEPALIVE=1 (Activates the KeepAlive mechanism):

    #############################
    http://h30499.www3.hp.com/t5/Data-Protector-Support-and-News/Data-Protector-Support-Tip-Linux-Only-Using-KeepAlive-to-resolve/td-p/6263015#.UuC5thQwc08
    #############################

    Kind regards,

    lex

Comments are closed.