Is this somebody else having the same problem? Just like with me, it started after upgrading beyond snv_104.
http://www.opensolaris.org/jive/thread.jspa?threadID=103071&tstart=0 On Wed, May 20, 2009 at 9:06 AM, Ross Smith <[email protected]> wrote: > Alan: Copying you in for info, but don't you dare reply while on holiday! > > > Hi guys, > > Well, we upgraded to snv_114 a couple of days ago (actually a clean > install, we just imported the zfs pool), but the CIFS crash problem is > still here. Overnight the server has stopped responding, and windows > clients are again returning: > > "c:\>net view \\rob-039v > System error 53 has occurred. > > The network path was not found." > > I've attached the full messages.txt file, the errors accessing the > server occurred this morning at around 8:25am. I've also attached the > procedure we use for re-installing the server. > > The key piece to me in the log appears to be this: > > May 20 08:09:39 ROB-039V idmap[4100]: [ID 696364 daemon.error] > Degraded operation (Global Catalog servers not > configured/discoverable). If you are running an SMB server in > workgroup mode, or if you're not running an SMB server, then you can > ignore this message > ... > May 20 08:23:47 ROB-039V tmpfs: [ID 518458 kern.warning] WARNING: > /var/run: File system full, swap space limit exceeded > May 20 08:24:55 ROB-039V smbd[4096]: [ID 266262 daemon.error] > ROBINSONS\Ross Smith: idmap failed > May 20 08:24:55 ROB-039V smbd[4096]: [ID 622271 daemon.error] > smb_idmap_restart: idmap_init failed (RPC error) > May 20 08:24:55 ROB-039V smbd[4096]: [ID 266262 daemon.error] > ROBINSONS\Build2003: idmap failed > May 20 08:24:56 ROB-039V smbd[4096]: [ID 266262 daemon.error] > ROBINSONS\Ross Smith: idmap failed > May 20 08:25:22 ROB-039V last message repeated 235 times > > Although if you search through the logs for idmap, you will see that > it quite regularly fails, it's just that most of the time it also > recovers: > > May 19 08:25:02 ROB-039V idmap[4100]: [ID 696364 daemon.error] > Degraded operation (Global Catalog servers not > configured/discoverable). If you are running an SMB server in > workgroup mode, or if you're not running an SMB server, then you can > ignore this message > ... > May 19 08:46:12 ROB-039V idmap[4100]: [ID 643943 daemon.notice] Normal > operation restored > > May 19 16:19:58 ROB-039V idmap[4100]: [ID 696364 daemon.error] > Degraded operation (Global Catalog servers not > configured/discoverable). If you are running an SMB server in > workgroup mode, or if you're not running an SMB server, then you can > ignore this message > ... > May 19 16:33:09 ROB-039V smbd[4096]: [ID 266262 daemon.error] > ROBINSONS\Build2003: idmap failed > May 19 16:34:44 ROB-039V last message repeated 8 times > May 19 16:35:14 ROB-039V idmap[4100]: [ID 643943 daemon.notice] Normal > operation restored > > May 20 00:14:50 ROB-039V idmap[4100]: [ID 696364 daemon.error] > Degraded operation (Global Catalog servers not > configured/discoverable). If you are running an SMB server in > workgroup mode, or if you're not running an SMB server, then you can > ignore this message > ... > May 20 00:53:18 ROB-039V idmap[4100]: [ID 643943 daemon.notice] Normal > operation restored > > When the CIFS server finally crashes, I'm not able to resume operation > by restarting either the idmap or the smb/server services, and > attempting to do so gives me this in the log: > > May 20 08:50:38 ROB-039V svc.startd[7]: [ID 122153 daemon.warning] > svc:/network/smb/server:default: Method or service exit timed out. > Killing contract 114. > May 20 08:57:06 ROB-039V last message repeated 387 times > May 20 08:57:07 ROB-039V svc.startd[7]: [ID 122153 daemon.warning] > svc:/network/smb/server:default: Method or service exit timed out. > Killing contract 114. > > I'm also not able to rejoin the domain, attempting to do so results in: > "failed to join ...: INTERNAL_ERROR". I assume because restarting the > service didn't work properly. > > After a reboot however things return to normal. > > The only unusual thing about the reboot is that attempting to browse > the share results in: > > "Z:\>net view \\rob-039v > System error 86 has occurred. > > The specified network password is not correct." > > This however is because Solaris on vmware looses the time on reboot, > and after a reboot I have to run: > # ntpdate rob-010 > # svcadm restart idmap > > Once this is done, the shares all work fine again. > > Ross > > > > On Sun, Apr 26, 2009 at 3:34 PM, Julian Pullen <[email protected]> wrote: >> From these messages there seams to be a problem of binding >> to the AD global catalog. I dont see an idmap GSS API (kerberos) >> errors to explain this. >> >> Regards >> >> Julian >> >> [email protected] wrote: >>> >>> I don't think it's smb that's hanging though, I think it's idmap. These >>> seem to be the key messages when it starts to fail: >>> >>> Feb 24 15:59:38 rob-039v idmap[10468]: [ID 326605 daemon.notice] Couldn't >>> open an LDAP connection to any global catalog server! >>> Feb 24 15:59:38 rob-039v last message repeated 2 times >>> Feb 24 15:59:38 rob-039v idmap[10468]: [ID 696364 daemon.error] Degraded >>> operation (failed to create batch for AD lookup). If you are running an SMB >>> server in workgroup mode, or if you're not running an SMB server, then you >>> can ignore this message >>> Feb 24 16:05:15 rob-039v smbd[457]: [ID 266262 daemon.error] >>> ROBINSONS\Ross Smith: idmap failed >>> >>> >>> >>> On Apr 23, 2009 8:17pm, Afshin Salek <[email protected]> wrote: >>> > Two fixes were pushed in snv_114: >>> > >>> > >>> > >>> > 6831172 smb_server.c is missing a few smb_server_release() calls >>> > >>> > 6831734 smbd exit on session reconnect with vc=0 >>> > >>> > >>> > >>> > The first one addresses an issue which could cause a hang upon restart. >>> > >>> > The second one addresses unexpected restarts. These may or may not >>> > >>> > address the problems here, but potentially they will. We either have >>> > >>> > to have more information to verify it or you have to wait for snv_114. >>> > >>> > >>> > >>> > Afshin >>> > >>> > >>> > >>> > [email protected] wrote: >>> > >>> > >>> > This is the original thread on the boards where I started discussing >>> it, I think the discussion has gotten fragmented though, so if anything >>> isn't clear please ask. >>> > >>> > >>> > >>> > My issue - CIFS keeps hanging in snv_106: >>> > >>> > http://www.opensolaris.org/jive/thread.jspa?messageID=346703䩏 >>> > >>> > >>> > >>> > Jays issue - CIFS dropping connection & corrupting files in snv_107: >>> > >>> > http://www.opensolaris.org/jive/thread.jspa?threadID=93479&tstart=0 >>> > >>> > >>> > >>> > Ross >>> > >>> > >>> > >>> > >>> > >>> > >>> > >>> > On Apr 23, 2009 7:13pm, Afshin Salek [email protected]> wrote: >>> > >>> > > Which daemon crashes, idmap or smbd? Any crash/core dumps? >>> > >>> > > >>> > >>> > > >>> > >>> > > >>> > >>> > > Afshin >>> > >>> > > >>> > >>> > > >>> > >>> > > >>> > >>> > > Ross Smith wrote: >>> > >>> > > >>> > >>> > > >>> > >>> > > Well, we upgraded to snv_112 yesterday to see if the idmap problems >>> > >>> > > >>> > >>> > > went away, but no such luck. It's failing almost daily, with >>> messages >>> > >>> > > >>> > >>> > > like this in /var/adm/messages: >>> > >>> > > >>> > >>> > > >>> > >>> > > >>> > >>> > > Apr 23 10:54:17 rob-039v idmap[4413]: [ID 696364 daemon.error] >>> > >>> > > >>> > >>> > > Degraded operation (Global Catalog servers not >>> > >>> > > >>> > >>> > > configured/discoverable). If you are running an SMB server in >>> > >>> > > >>> > >>> > > workgroup mode, or if you're not running an SMB server, then you can >>> > >>> > > >>> > >>> > > ignore this message >>> > >>> > > >>> > >>> > > >>> > >>> > > >>> > >>> > > This was absolutely fine in snv_94, but crashes in clean installs of >>> > >>> > > >>> > >>> > > OpenSolaris 2008.11 (101b), and snv_112. I'm having to constantly >>> > >>> > > >>> > >>> > > reboot this server because the CIFS shares keep going unavailable to >>> > >>> > > >>> > >>> > > our clients, could somebody please let me know if it's being looked >>> > >>> > > >>> > >>> > > into? >>> > >>> > > >>> > >>> > > >>> > >>> > > >>> > >>> > > thanks, >>> > >>> > > >>> > >>> > > >>> > >>> > > >>> > >>> > > Ross >>> > >>> > > >>> > >>> > > >>> > >>> > > >>> > >>> > > >>> > >>> > > >>> > >>> > > >>> > >>> > > >>> > >>> > > >>> > >>> > > >>> > >>> > > On Thu, Feb 26, 2009 at 11:02 AM, Alan.M.Wright [email protected]> wrote: >>> > >>> > > >>> > >>> > > >>> > >>> > > Julian, >>> > >>> > > >>> > >>> > > >>> > >>> > > >>> > >>> > > Some idmap things in here. Are either of these issues familiar to >>> you? >>> > >>> > > >>> > >>> > > >>> > >>> > > >>> > >>> > > I hadn't seen these messages. Perhaps the forum forwarding is >>> messed >>> > >>> > > >>> > >>> > > up again, I'll notify the admins. >>> > >>> > > >>> > >>> > > >>> > >>> > > >>> > >>> > > Alan >>> > >>> > > >>> > >>> > > >>> > >>> > > >>> > >>> > > ----- Original Message ----- From: "Ross Smith" >>> [email protected]> >>> > >>> > > >>> > >>> > > To: "Alan M Wright" [email protected]>; "Afshin Salek" >>> > >>> > > >>> > >>> > > Sent: Thursday, February 26, 2009 1:44 AM >>> > >>> > > >>> > >>> > > Subject: Problems with cifs in snv_106 >>> > >>> > > >>> > >>> > > >>> > >>> > > >>> > >>> > > >>> > >>> > > >>> > >>> > > >>> > >>> > > Hi guys, >>> > >>> > > >>> > >>> > > >>> > >>> > > >>> > >>> > > Sorry to mail directly but I'm having real problems with CIFS since >>> > >>> > > >>> > >>> > > upgrading to 106, and there are problems being reported with 107 >>> too, >>> > >>> > > >>> > >>> > > and I've not seen anybody from Sun responding to these on the >>> forums. >>> > >>> > > >>> > >>> > > >>> > >>> > > >>> > >>> > > My issue - CIFS keeps hanging in snv_106: >>> > >>> > > >>> > >>> > > http://www.opensolaris.org/jive/thread.jspa?messageID=346703䩏 >>> > >>> > > >>> > >>> > > >>> > >>> > > >>> > >>> > > Jays issue - CIFS dropping connection & corrupting files in snv_107: >>> > >>> > > >>> > >>> > > http://www.opensolaris.org/jive/thread.jspa?threadID=93479&tstart=0 >>> > >>> > > >>> > >>> > > >>> > >>> > > >>> > >>> > > Would be good to hear from you guys on the forum, even if it's just >>> to >>> > >>> > > >>> > >>> > > say you're aware of the issues. >>> > >>> > > >>> > >>> > > >>> > >>> > > >>> > >>> > > thanks, >>> > >>> > > >>> > >>> > > >>> > >>> > > >>> > >>> > > Ross >>> > >>> > > >>> > >>> > > >>> > >>> > > >>> > >>> > > >>> > >>> > >> > _______________________________________________ cifs-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/cifs-discuss
