|To:||Dave Chinner <david@xxxxxxxxxxxxx>|
|Subject:||Re: 2.6.31+126.96.36.199: XFS - All I/O locks up to D-state after 24-48 hours (sysrq-t+w available)|
|From:||Justin Piszcz <jpiszcz@xxxxxxxxxxxxxxx>|
|Date:||Mon, 2 Nov 2009 16:46:10 -0500 (EST)|
|Cc:||linux-kernel@xxxxxxxxxxxxxxx, linux-raid@xxxxxxxxxxxxxxx, xfs@xxxxxxxxxxx, Alan Piszcz <ap@xxxxxxxxxxxxx>|
|References:||<alpine.DEB.2.00.0910171825270.16781@xxxxxxxxxxxxxxxx> <alpine.DEB.2.00.0910181607040.27363@xxxxxxxxxxxxxxxx> <20091019030456.GS9464@xxxxxxxxxxxxxxxx> <alpine.DEB.2.00.0910190431180.23395@xxxxxxxxxxxxxxxx> <20091020003358.GW9464@xxxxxxxxxxxxxxxx> <alpine.DEB.2.00.0910200431290.21878@xxxxxxxxxxxxxxxx> <alpine.DEB.2.00.0910210618210.10288@xxxxxxxxxxxxxxxx> <alpine.DEB.2.00.0910221849001.24576@xxxxxxxxxxxxxxxx> <alpine.DEB.2.00.0910260723150.22950@xxxxxxxxxxxxxxxx>|
|User-agent:||Alpine 2.00 (DEB 1167 2008-08-23)|
On Mon, 26 Oct 2009, Justin Piszcz wrote:
On Thu, 22 Oct 2009, Justin Piszcz wrote:Any other ideas?Currently stuck on 188.8.131.52.. (no issues, no lockups)-- Box normally has no load at all either.. Has anyone else reported similar problems?Justin.--Currently running 2.6.31-rc1 for 2 days now, no crashes, will go to -rc2 later today and wait another 48 hours.Justin.
Kernel report: 2.6.31-rc1: no crash - uptime 2+ days 2.6.31-rc2: no crash - uptime 2+ days 2.6.31-rc3: no crash - uptime 2+ days 2.6.31-rc4: no crash but network kept dropping out 2.6.31-rc5: cannot test, service owner needs host available 2.6.31-rc6: cannot test, service owner needs host available 2.6.31-rc7: cannot test, service owner needs host available 2.6.31-rc8: cannot test, service owner needs host available 2.6.31-rc9: cannot test, service owner needs host available 2.6.31.x: locks up D-state It would be somewhere between -rc4 and 2.6.31.x. Justin.
|<Prev in Thread]||Current Thread||[Next in Thread>|
|Previous by Date:||Re: XFS and DPX files, Eric Sandeen|
|Next by Date:||Re: XFS and DPX files, AndrewL733@xxxxxxx|
|Previous by Thread:||RE: [PATCH] xfs: fix mmap_sem vs iolock lock order inversion inxfs_free_eofblocks, Alex Elder|
|Next by Thread:||RE: [PATCH] xfs: reset the i_iolock lock class in the reclaim path, Alex Elder|
|Indexes:||[Date] [Thread] [Top] [All Lists]|