toasters

toasters@lists.teaparty.net

1 participants
13523 discussions

How long can Windows survive loss of storage?
by Stuart Kendrick 26 Jan '12

26 Jan '12

Hi folks, We have MS SQL Server (a handful of instances, riding on top of MS Cluster Services) attached via Fibre Channel to a filer. Whenever we perform a takeover, SQL Server crumps. The DBAs shut it down, restart, repair/restore their databases (unhappy times), life returns to normal. We've replicated this behavior a handful of times: every single time we experience a takeover, whether administratively or Murphy-induced. I'd like to think that we can do better than this. This particular filer services five clients: the two MS SQL Servers via Fibre Channel and three Exchange servers via iSCSI. The iSCSI clients notice the takeover, but their initiators do the multipath thing and ride through the event fine. In the log extract below (heavily edited), we can see the iSCSI and FCP services going down at 21:47:25. And coming back on-line at 21:47:40 (FCP) and 21:47:49 (iSCSI). The iSCSI client 'hamlet' re-establishes its session at 21:47:51. The rest of the log extract records the downward spiral of the Windows server 'custard-1'. To my way of thinking, iSCSI was unavailable for 24 seconds, FCP for 15 seconds. (1) Are the iSCSI and FCP services really and truly available once the 'fcp.service.startup' and 'iscsi.service.startup' messages are logged, or might there be a delay before storage is actually available to clients? (2) How could we shrink takeover times? (3) Why does MS Windows over FC have such trouble surviving a 15 second disruption to storage? (4) How could we increase MS Windows timeouts? (5) How long can /your/ MS Windows FCP clients survive an interruption to storage? Toaster: FAS3020 running 7.2.4 SQL Server Clients: Windows 2008 running SnapDrive & SnapManager for SQL Server Exchange Clients: Windows 2003 running SnapDrive & SnapManager for Exchange Log extract below --sk Stuart Kendrick FHCRC Feb 7 21:47:22 toaster-b-svif2 [toaster-b: cf.misc.operatorTakeover:warning]: Cluster monitor: takeover initiated by operator Feb 7 21:47:22 toaster-b-svif2 [toaster-b: cf.fsm.nfo.acceptTakeoverReq:warning]: Negotiated failover: accepting takeover request by partner, reason: operator initiated cf takeover. Asking partner to shutdown gracefully; will takeover in at most 180 seconds. Feb 7 21:47:25 toaster-a-svif2 [toaster-a: cf.fsm.nfo.startingGracefulShutdown:warning]: Negotiated failover: starting graceful shutdown. Feb 7 21:47:25 toaster-a-svif2 [toaster-a: kern.shutdown:notice]: System shut down because : "reboot". Feb 7 21:47:25 toaster-a-svif2 [toaster-a: iscsi.service.shutdown:info]: iSCSI service shutdown Feb 7 21:47:25 toaster-a-svif2 [toaster-a: fcp.service.shutdown:info]: FCP service shutdown Feb 7 21:47:32 toaster-b-svif2 [toaster-b: cf.fsm.firmwareStatus:info]: Cluster monitor: partner rebooting Feb 7 21:47:32 toaster-b-svif2 [toaster-b: cf.fsm.nfo.partnerShutdown:warning]: Negotiated failover: partner has shutdown Feb 7 21:47:32 toaster-b-svif2 [toaster-b: cf.fsm.takeover.nfo:info]: Cluster monitor: takeover attempted after 'cf takeover'. command Feb 7 21:47:32 toaster-b-svif2 [toaster-b: cf.fsm.stateTransit:warning]: Cluster monitor: UP --> TAKEOVER Feb 7 21:47:32 toaster-b-svif2 [toaster-b: cf.fm.takeoverStarted:warning]: Cluster monitor: takeover started Feb 7 21:47:34 toaster-b-svif2 [toaster-b: cf_takeover:info]: NVRAM takeover: partner nvram is disabled Feb 7 21:47:39 toaster-b-svif2 [toaster-a/toaster-b: wafl.takeover.nvram.missing:error]: WAFL takeover: no partner area found during wafl replay Feb 7 21:47:40 toaster-b-svif2 [toaster-a/toaster-b: wafl.replay.done:info]: WAFL log replay completed, 0 seconds Feb 7 21:47:40 toaster-b-svif2 [toaster-a/toaster-b: fcp.service.startup:info]: FCP service startup 1 sys Error custard-1 2010-02-07 21:47:44 61203 ontapdsm DSM ID 0400010d has transitioned to the failed state. 1 sys Error custard-1 2010-02-07 21:47:44 61203 ontapdsm DSM ID 04000107 has transitioned to the failed state. 1 sys Error custard-1 2010-02-07 21:47:44 61203 ontapdsm DSM ID 0400010b has transitioned to the failed state. 1 sys Error custard-1 2010-02-07 21:47:44 61203 ontapdsm DSM ID 0300010b has transitioned to the failed state. [...] 1 sys Error custard-1 2010-02-07 21:47:44 61142 ontapdsm Nexus ID 03000101 has failed. Feb 7 21:47:48 toaster-b-svif2 [toaster-a/toaster-b: cf.fm.takeoverDetectionSeconds.Default:warning]: option cf.takeover.detection.seconds is set to 10 seconds which is below the NetApp advised value of 15 seconds. False takeovers and/or takeovers without diagnostic core-dumps might occur. Feb 7 21:47:49 toaster-b-svif2 [toaster-a/toaster-b: iscsi.service.startup:info]: iSCSI service startup Feb 7 21:47:50 toaster-b-svif2 [toaster-b (takeover): cf.rsrc.transitTime:notice]: Top Takeover transit times wafl_restart=7400 {vdisk=7214, restarters=186}, wafl=5049, registry_prerc=613, registry_postrc_phase2=555, rc=553 {ifconfig=99, ifconfig=93, hostname=57, options=41, options=27, options=26, route=2, vif=1, vif=1, vif=1}, raid=525, registry_postrc_phase1=403, wafl_sync=241, raid_replay=221, cifs=216 Feb 7 21:47:50 toaster-b-svif2 [toaster-b (takeover): cf.fm.takeoverComplete:warning]: Cluster monitor: takeover completed Feb 7 21:47:50 toaster-b-svif2 [toaster-b (takeover): cf.fm.takeoverDuration:warning]: Cluster monitor: takeover duration time is 18 seconds Feb 7 21:47:51 toaster-b-svif2 [toaster-a/toaster-b: iscsi.notice:notice]: ISCSI: New session from initiator iqn.1990-04.org.fhcrc:hamlet.fhcrc.org at IP addr 10.111.152.44 Feb 7 21:47:51 toaster-a-svif1 [toaster-a/toaster-b: iscsi.notice:notice]: ISCSI: New session from initiator iqn.1990-04.org.fhcrc:hamlet.fhcrc.org at IP addr 10.111.152.44 1 sys Warning custard-1 2010-02-07 21:47:57 61051 ontapdsm DSM ID 03000100 failed path verification. 1 sys Warning custard-1 2010-02-07 21:47:57 61051 ontapdsm DSM ID 03000101 failed path verification. 1 sys Warning custard-1 2010-02-07 21:47:57 61051 ontapdsm DSM ID 03000102 failed path verification. 1 sys Warning custard-1 2010-02-07 21:47:57 61077 ontapdsm DSM ID 03000111 has initiated a fail-over. [...] 1 sys Error custard-1 2010-02-07 21:47:57 61124 ontapdsm The port servicing DSM ID 03000111 reported the logical unit did not respond to selection. [...] Feb 7 21:48:01 toaster-b-svif2 [toaster-b (takeover): monitor.globalStatus.critical:CRITICAL]: This node has taken over toaster-a. Feb 7 21:48:01 toaster-b-svif2 [toaster-a/toaster-b: monitor.globalStatus.critical:CRITICAL]: toaster-b has taken over this node. Feb 7 21:48:01 toaster-a-svif1 [toaster-a/toaster-b: monitor.globalStatus.critical:CRITICAL]: toaster-b has taken over this node. 1 sys Warning custard-1 2010-02-07 21:48:04 61204 ontapdsm DSM ID 0400010f is in the process of being removed. 1 sys Warning custard-1 2010-02-07 21:48:04 61205 ontapdsm DSM ID 0400010f was removed. 1 sys Warning custard-1 2010-02-07 21:48:04 61205 ontapdsm DSM ID 0400010e was removed. [...] 1 sys Error custard-1 2010-02-07 21:48:04 16 mpio A fail-over on DeviceMPIODisk108 occurred. [...] 1 sys Warning custard-1 2010-02-07 21:48:04 17 mpio DeviceMPIODisk108 is currently in a degraded state. One or more paths have failed though the process is now complete. Feb 7 21:48:09 toaster-b-svif2 [toaster-a/toaster-b: syslogd:info]: syslogd: restarted Feb 7 21:48:09 toaster-a-svif1 [toaster-a/toaster-b: syslogd:info]: syslogd: restarted Feb 7 21:48:11 toaster-b-svif2 [toaster-a/toaster-b: nbt.nbns.registrationComplete:info]: NBT: All CIFS name registrations have completed for the partner server. Feb 7 21:48:11 toaster-a-svif1 [toaster-a/toaster-b: nbt.nbns.registrationComplete:info]: NBT: All CIFS name registrations have completed for the partner server. 1 sys Warning custard-1 2010-02-07 21:48:24 129 ql2300 Reset to device DeviceRaidPort1 was issued. 1 sys Warning custard-1 2010-02-07 21:48:28 1123 ClusSvc The node lost communication with cluster node 'custard-2' on network 'Public (Team)'. 1 sys Warning custard-1 2010-02-07 21:48:28 1123 ClusSvc The node lost communication with cluster node 'custard-2' on network 'Private 172.16 (PCI)'. 1 sys Warning custard-1 2010-02-07 21:48:28 1123 ClusSvc The node lost communication with cluster node 'custard-2' on network 'Private 192.168 (Onboard)'. 1 sys Error custard-1 2010-02-07 21:48:43 1209 ClusDisk Cluster service is requesting a bus reset for device DeviceClusDisk0. Feb 7 21:48:46 toaster-b-svif2 [toaster-b (takeover): scsitarget.ispfct.targetReset:notice]: FCP Target 0b: Target was Reset by the Initiator at Port Id: 0x10000 (WWPN 2100001b3201ba33) 1 sys Error custard-1 2010-02-07 21:48:51 1209 ClusDisk Cluster service is requesting a bus reset for device DeviceClusDisk0. 1 sys Error custard-1 2010-02-07 21:48:53 1118 ClusNet Cluster service was terminated as requested by Node 2. 1 sys Error custard-1 2010-02-07 21:48:53 7031 Service Control Manager The Cluster Service service terminated unexpectedly. It has done this 1 time(s). The following corrective action will be taken in 60000 milliseconds: Restart the service. 5 sys Error custard-1 2010-02-07 21:48:54 1036 ClusSvc Cluster disk resource '' did not respond to a SCSI maintenance command. 1 sys Error custard-1 2010-02-07 21:48:54 1215 ClusSvc Cluster Network Name custard is no longer registered with its hosting system. The associated resource name is ''. 1 sys Error custard-1 2010-02-07 21:48:54 1077 ClusSvc The TCP/IP interface for Cluster IP Address '' has failed. 1 sys Warning custard-1 2010-02-07 21:48:56 50 Ntfs {Delayed Write Failed} Windows was unable to save all the data for the file . The data has been lost. This error may be caused by a failure of your computer hardware or network connection. Please try to save this file elsewhere. 1 sys Warning custard-1 2010-02-07 21:48:57 57 Ftdisk The system failed to flush data to the transaction log. Corruption may occur.

6 7

maximum NDMP sessions?
by Randy Rue 26 Jan '12

26 Jan '12

Hello All, We've previously been under the impression that our v3170 could only run 16 NDMP sessions at a time. But our reporting seems to show as many as 27 running. Did this change in our last upgrade? We're currently at 7.3.5.1P5 Randy

3 3

MS Cluster clients locked out of LUNs
by Stuart Kendrick 24 Jan '12

24 Jan '12

Hi folks, After a takeover, we find that some (though not all) LUNs allocated to Microsoft Clustering clients show up in the Microsoft GUI as 'offline' or 'failed'. Rebooting the clients doesn't help. When we disable the Cluster disk drivers (make the client a simple iSCSI client), we can connect, read & write, etc. On the filer side, no sign of a reservation. We've replicated this behavior on three occasions over the last year (every time we experience a takeover, either administratively induced or accident induced). The clients run SnapDrive & SnapManager for SQL Server. To resolve this, we copy our data off (while connected via the vanilla iSCSI client), destroy the LUNs, rebuild, copy data back. Alternatively, we can wait a few days (not sure how many), whereupon the issue resolves itself (we've seen this once ... test environment). Anyone else seen this? Insights? What might be happening? NetApp v3140 running 7.3.5.1P5 Spinning rust provided by 3Par T800 Clients running Win2008R2 --sk Stuart Kendrick FHCRC

2 1

read only role?
by Randy Rue 24 Jan '12

24 Jan '12

Hello All, Looking to add a read-only role for techs in our department who are investigating some things but don't need (or want) to risk breaking stuff. Is there a "standard" list of capabilities that can be added to a role that will give the ability to see stuff but not break stuff? Thanks in advance, Randy Rue

5 7

Massive messages files.
by Jeff Cleverley 20 Jan '12

20 Jan '12

Greetings, I upgraded to 7.3.5.1P4 from 7.3.3P5 on January 1st. Since then my messages log files have gotten huge. My current one is 123 MB and the average size in December was 250K. I've always had the syslog.conf set up to send info levels to the messages file. Since the upgrade I've been getting a log of messages about things that used to never show up such as MNTPool and ExportPool entries. There are other entries that never showed up before either. I like having the info level because it shows me who is connecting via cifs, time stamp update, nis updates, etc. I would rather not have to change all the logging to the warning level. Parsing through messages files with 800k lines in it is not pretty :-) Is there another way to tune what shows up at the info level in 7.3.5? Thanks, Jeff -- Jeff Cleverley Unix Systems Administrator 4380 Ziegler Road Fort Collins, Colorado 80525 970-288-4611

2 2

v-filers and external network isolation
by Craig A. Falls 19 Jan '12

19 Jan '12

Good People, I am a long time reader, novice poster, and I have a question for you. I have a environment where I want to put a single physical filer across two distinct networks (one internal and once external). Using v-filers I want to provide services for both networks. Generally I would use two physical filers for this, so attacks on the physical filer in the external network would only compromise storage in the external network. In using a single physical filer, and using v-filers, I am hoping to achieve the same results, the issue is however that I can still get to v-filer0, and thus the filer as a whole from the IP address on the interface that is hosting the v-filers provisioned to the external network. Thus providing an attack vector to something that before (with two physical filers) was completely inaccessible. I see two possible solutions that I am not sure are implementable: (1) Somehow creating the interface connected to the external network without an IP address, thus v-filers would have IP address on that interface, and be accessible, but the v-filer0 would not be accessible, as there is no IP to access it, or (2) Somehow deny ssh to v-filer0 from the external network interface, but it would still have to work for the v-filers on that interface. Not sure if these are possible, or an extensive list of the ways to achieve what I am attempting. Also in re-reading this SSH isn't the only attach vector, http and any other management interface would have to be denied from the external interfaces as well. Any advice or thoughts would be much appreciated. thanks c

4 3

Head swap
by Simon Ogilvie 19 Jan '12

19 Jan '12

Hi all, I have a query regarding a head-swap upgrade which I've never attempted before, so any help would be much appreciated! We currently have a single-head FAS3020 running a small number of SATA and FC shelves. The head is EOL and so I've been able to persuade my bean counters to buy me a pair of FAS3240 heads to use as an HA pair. We're keeping all the disk shelves, so I "just" need to swap the heads over. From what I've read, it's best practice to upgrade the old head to the version of ONTAP that the new head will run, but we don't have that option as the 3020 won't run ONTAP 8 and the 3240's won't run ONTAP 7! This is our central storage system, using CIFS, NFS and LUNs over FC, so getting it wrong isn't really an option. Is this going to be achievable or should I call professional services in? I'd rather do this myself if possible, not least because PS are rather expensive... Can I just copy the latest ONTAP onto the existing vol0 and shut the head down, power up the new head and make the disks owned by the new head? I assume I'll also have to tell the FC switches about the new WWPNs for the LUNs to work. Will that work or will I end up with a very large, very expensive brick? Thanks! Simon.

15 19

cifs operations
by Crawford, Mark (CBC) 15 Jan '12

15 Jan '12

Hello toasters: Can anyone explain what CIFS operations that are neither read nor write operations? In addition, can anyone think of what process(es) could be causing these? ops/s reads(n, KB/s) writes(n, KB/s) suspect/s 860 | 0 0 | 0 0 | 0 | We have 1500+ windows 7 workstations connecting to their homeshares on this head. Lately we have been seeing CIFS latency where the user workstations are hanging within their windows explorer window. When this occurs, a cifs top shows high "ops/s" with no associated reads or writes. We are running DATA ONTap v7.3.5.1P1. Any insights would be greatly appreciated. Confidentiality Notice: This e-mail is intended only for the personal and confidential use of the individual to whom it is addressed and may contain information that is privileged, confidential and protected by law. If you are not the intended recipient, you are hereby notified that any use or disclosure of this information is strictly prohibited. If you have received this message in error, please notify the sender immediately by reply e-mail and delete the original message. Your compliance is appreciated.

3 2

Re: cifs operations
by Page, Jeremy 15 Jan '12

15 Jan '12

If your clients are set to index shares this can happen. ----- Reply message ----- From: "Webster, Stetson" <Stetson.Webster(a)netapp.com> To: "Crawford, Mark (CBC)" <Mark.Crawford(a)CapBlueCross.COM> Cc: "toasters(a)teaparty.net" <toasters(a)teaparty.net> Subject: cifs operations Date: Sun, Jan 15, 2012 9:10 am In short, attribute lookups and directory listings. In long, it's hard to tell the root cause without seeing the big picture within the auspices of a support case. For example, what you describe could simply be a symptom of a broader issue or even a mere client configuration change. It's hard to tell, but hopefully you have enough to get started. Sent from my iThumbs. Please pardon my thumbmanship. On Jan 15, 2012, at 8:34 AM, "Crawford, Mark (CBC)" <Mark.Crawford(a)CapBlueCross.COM<mailto:Mark.Crawford@CapBlueCross.COM>> wrote: Hello toasters: Can anyone explain what CIFS operations that are neither read nor write operations? In addition, can anyone think of what process(es) could be causing these? ops/s reads(n, KB/s) writes(n, KB/s) suspect/s 860 | 0 0 | 0 0 | 0 | We have 1500+ windows 7 workstations connecting to their homeshares on this head. Lately we have been seeing CIFS latency where the user workstations are hanging within their windows explorer window. When this occurs, a cifs top shows high “ops/s” with no associated reads or writes. We are running DATA ONTap v7.3.5.1P1. Any insights would be greatly appreciated. Confidentiality Notice: This e-mail is intended only for the personal and confidential use of the individual to whom it is addressed and may contain information that is privileged, confidential and protected by law. If you are not the intended recipient, you are hereby notified that any use or disclosure of this information is strictly prohibited. If you have received this message in error, please notify the sender immediately by reply e-mail and delete the original message. Your compliance is appreciated. _______________________________________________ Toasters mailing list Toasters(a)teaparty.net<mailto:Toasters@teaparty.net> http://www.teaparty.net/mailman/listinfo/toasters Please be advised that this email may contain confidential information. If you are not the intended recipient, please notify us by email by replying to the sender and delete this message. The sender disclaims that the content of this email constitutes an offer to enter into, or the acceptance of, any agreement; provided that the foregoing does not invalidate the binding effect of any digital or other electronic reproduction of a manual signature that is included in any attachment.

1 0

DS14MK4 as DAS
by Don Elmore 08 Jan '12

08 Jan '12

Apologies if this has been asked before. We have a DS14MK4 shelf with 144G FC disks and 2 ESH4 modules that's no longer in production. I've heard of people using these shelves as direct-attached storage (no filer head, just directly cabled to a host), but I haven't been able to pull up any details after searching online. I figure if anybody has tried something this silly, they're probably on this mailing list :) Has anybody been able to get this to work? Any tips/tricks? Thanks! Don Elmore IT Manager Volition, Inc. ________________________________ This message, including any attachments, may contain privileged and/or confidential information. Any distribution or use of this email by anyone other than the intended recipient(s) is strictly prohibited. If you are not the intended recipient, please notify the sender immediately and delete all copies. Thank you.

5 5

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

toasters