toasters

toasters@lists.teaparty.net

13520 discussions

filer CPU usage - getting warmer
by Ambrose_Earle＠shamrockfoods.com 13 Jul '02

13 Jul '02

Armed with the above information, the next step was to perform some packet traces to see what I could find. Thankfully (?) the problem is continuous, so I could time my traces to only capture the spikes and I could wait until late in the night to limit the extraneous packets as much as possible. Doing this a bunch of times I was able to isolate what seems to be the problem. However, correlation does not imply causation, it merely points a finger in its direction. I won't really know until Monday when I'm able to stop the process to see if the spikes disappear. Because the traffic flows over a vif, I can only capture the entire vif and not an individual interface. I also do not know how our 6509 was configured, so I can't attempt to calculate over which port the traffic from this client would be directed, but the amount and type of traffic leads me to believe that I have found the culprit. This particular application server runs a program within a DOS on a Windows 98 PC. It sends hundreds of search requests at exactly the same time as the CPU spikes. Researching this shows some related bugs - 35047 or 40348. If this is the problem, then I'll see what can be done about the application. If nothing, then I'll be opening a case. Thanks to the personal replies I received from NetApps.

2 1

filer CPU usage - more information?
by Ambrose_Earle＠shamrockfoods.com 13 Jul '02

13 Jul '02

After digging deeper, it seems that the spikes correlate perfectly with the "driver" process for e3c. e3c is part of a four-way trunk group running to a Cisco 6509. Any thoughts about what might make one NIC need so much attention? I have interleaved the ps commands about where they occurred while a sysstat was running. I hope this is clear F840C*> ps -z CPU NFS CIFS HTTP Total Net kB/s Disk kB/s Tape kB/s Cache Cache CP CP Disk in out read write read write age hit time ty util 1% 88 52 0 140 237 40 24 76 0 0 27 100% 7% : 4% 6% 49 202 0 251 57 408 3173 0 0 0 27 98% 0% - 38% 3% 33 253 0 286 205 1030 684 16 0 0 27 99% 0% - 9% 4% 89 188 0 277 152 506 4188 0 0 0 27 95% 0% - 41% 2% 127 144 0 271 151 948 837 0 0 0 27 99% 0% - 14% 3% 119 2 0 121 102 32 32 16 0 0 27 100% 0% - 1% 0% 24 0 0 24 123 28 0 0 0 0 27 100% 0% - 0% 1% 28 0 0 28 18 2 0 0 0 0 27 100% 0% - 0% 3% 116 137 0 253 98 250 2664 16 0 0 27 96% 0% - 20% 3% 35 160 0 195 184 1991 2300 0 0 0 27 98% 0% - 28% F840C*> ps -h Process statistics over 9.799 seconds... ID State Domain %CPU StackUsed %StackUsed Name 1 RR i 97% 276 6% idle_thread0 F840C*> ps -z 6% 25 114 0 139 14 58 1976 2240 0 0 27 98% 27% T 35% 3% 33 82 0 115 106 967 3916 16 0 0 27 95% 0% - 44% 4% 141 78 0 219 137 1005 4392 0 0 0 27 92% 0% - 42% 7% 105 74 0 179 53 700 5220 0 0 0 27 96% 0% - 51% 5% 56 148 0 204 317 1515 4292 16 0 0 27 97% 0% - 45% 6% 82 217 0 299 451 1408 4888 0 0 0 27 97% 0% - 39% 4% 164 163 0 327 86 2053 3628 0 0 0 27 98% 0% - 36% 1% 62 5 0 67 107 148 32 16 0 0 27 100% 0% - 1% F840C*> ps -h Process statistics over 7.586 seconds... ID State Domain %CPU StackUsed %StackUsed Name 1 RR i 98% 276 6% idle_thread0 0% 17 0 0 17 107 4 0 0 0 0 27 100% 0% - 0% 0% 20 0 0 20 1 1 0 0 0 0 27 100% 0% - 0% 1% 110 1 0 111 70 11 32 16 0 0 27 100% 0% - 1% 6% 39 165 0 204 107 78 1740 2128 0 0 27 98% 21% T 27% 8% 22 409 0 431 177 1691 5345 0 0 0 27 97% 0% - 35% F840C*> ps -z 4% 75 320 0 395 136 2526 2828 16 0 0 27 98% 0% - 20% 46% 104 404 0 508 179 432 0 0 0 0 27 100% 0% - 0% 94% 360 313 0 673 143 157 0 0 0 0 27 100% 0% - 0% 92% 34 311 0 345 131 86 40 16 0 0 27 100% 0% - 2% 48% 14 589 0 603 366 815 0 0 0 0 27 100% 0% - 0% 5% 209 633 0 842 205 1985 732 0 0 0 27 100% 0% - 15% F840C*> ps -h Process statistics over 5.988 seconds... ID State Domain %CPU StackUsed %StackUsed Name 1 RR i 51% 276 6% idle_thread0 17 BR k 18% 9060 27% FastEnet-10/100/e3c 8% 134 184 0 318 369 696 4463 16 0 0 27 97% 0% - 28% 1% 143 58 0 201 272 1871 156 0 0 0 27 97% 0% - 1% F840C*> ps -h Process statistics over 386.288 seconds... ID State Domain %CPU StackUsed %StackUsed Name 1 RR i 84% 276 6% idle_thread0 17 BR k 5% 9060 27% FastEnet-10/100/e3c

1 0

filer CPU usage
by Ambrose_Earle＠shamrockfoods.com 12 Jul '02

12 Jul '02

I'm trying to track down a repeated CPU spike on one of my filers. About every 30 seconds, sysstat will show the CPU usage at or near 100%. There is no pattern to this other than the length of time between spikes. What are some common things I should look for? What commands might help me track down the culprit? Here's an example: 96% 86 733 0 819 378 1054 36 0 0 0 30 100% 0% - 2% 95% 62 609 0 671 357 819 1106 16 0 0 30 100% 0% - 15% 32% 80 597 0 677 532 363 1352 0 0 0 30 100% 0% - 17% 9% 47 750 0 797 1209 715 2960 0 0 0 30 96% 0% - 37% 11% 36 777 0 813 1310 1597 5564 16 0 0 30 97% 0% - 55% 5% 102 876 0 978 940 1837 296 0 0 0 30 99% 0% - 11% 2% 59 165 0 224 370 403 16 0 0 0 30 100% 0% - 1% 15% 27 504 0 531 53 228 744 7108 0 0 30 100% 29% T 25% 11% 45 827 0 872 1222 559 56 68 0 0 30 100% 7% : 4% 1% 81 102 0 183 367 764 0 0 0 0 30 100% 0% - 0% 1% 39 79 0 118 38 166 32 16 0 0 30 100% 0% - 2% 3% 32 345 0 377 128 216 768 0 0 0 30 96% 0% - 15% 9% 159 756 0 915 310 1122 6048 0 0 0 30 98% 0% - 72% 10% 371 543 0 914 1764 2896 2660 16 0 0 30 98% 0% - 34% 1% 127 85 0 212 1122 729 16 0 0 0 30 99% 0% - 1% 1% 24 67 0 91 183 179 4 0 0 0 30 100% 0% - 1% 2% 248 101 0 349 69 177 56 16 0 0 30 100% 0% - 2% 2% 144 104 0 248 100 476 0 0 0 0 30 100% 0% - 0% 18% 18 156 0 174 383 297 2345 5569 0 0 30 100% 30% T 29% 5% 38 350 0 388 102 239 1376 2300 0 0 30 100% 18% : 21% 6% 374 262 0 636 148 582 3504 0 0 0 30 94% 0% - 39% 5% 80 236 0 316 231 1251 4724 0 0 0 30 94% 0% - 41% 4% 0 266 0 266 396 2253 3128 16 0 0 30 97% 0% - 27% 1% 37 97 0 134 245 609 0 0 0 0 30 100% 0% - 0% 1% 53 83 0 136 155 275 0 0 0 0 30 100% 0% - 0% 1% 76 82 0 158 57 151 48 16 0 0 30 100% 0% - 1% 2% 57 169 0 226 38 343 8 0 0 0 30 100% 0% - 1% 2% 179 167 0 346 170 311 28 0 0 0 30 99% 0% - 3% 8% 50 310 0 360 205 670 6068 16 0 0 30 97% 0% - 54% 11% 130 406 0 536 530 1192 4612 3260 0 0 30 98% 31% T 60% 8% 27 768 0 795 184 1086 3748 92 0 0 30 98% 10% : 35% 43% 64 500 0 564 190 949 3836 16 0 0 30 100% 0% - 21% 92% 46 677 0 723 1178 1455 3383 0 0 0 30 100% 0% - 21%

1 0

SNMP Support for Filers
by Ben Rockwood 11 Jul '02

11 Jul '02

Hello Toasters, I am working on some new apps for monitoring and reporting on my Filers, however I'm running into a roadblock because I'm only using SNMPv1. Does OnTap support SNMPv2c? If so, how do I enable it? I can't seem to find any options for it. Thank You. Ben Rockwood

1 0

RE: How can I test or simulate an SNMP trap?
by Pero, Eric P (ETSD, IT) 10 Jul '02

10 Jul '02

Thanks for the feedback. I am running 6.1.3R1. Since the 6.2 solution is out, how do I parse my logs and forward the events to a SNMP host? Eric -----Original Message----- From: Chris Thompson [mailto:cet1@cus.cam.ac.uk] Sent: Wednesday, July 10, 2002 12:51 PM To: Drew(a)cooperneff.com; Eric.Pero(a)thehartford.com Cc: toasters(a)mathworks.com Subject: Re: How can I test or simulate an SNMP trap? Eric.Pero(a)thehartford.com (Eric Pero) wrote: > > How do you enable the trap for QTREE quotas (or is this already on when you > do the "snmp init 1" with the other factory traps)? Drew(a)cooperneff.com (Drew O'Donnell) replies: < < Traps are only generated when the volume is full, not the qtree. We parse < our logs because messages picks up a qtree error when the quota is exceeded, < then we forward this to openview. If you are running ONTAP 6.2 or above, you should be able to set a soft quota for a qtree (or user, or group, for that matter) that will among other things trigger an SNMP trap when exceeded. Or so the documentation says: I haven't actually tried this. Chris Thompson Email: cet1(a)cam.ac.uk This communication, including attachments, is for the exclusive use of addressee and may contain proprietary, confidential or privileged information. If you are not the intended recipient, any use, copying, disclosure, dissemination or distribution is strictly prohibited. If you are not the intended recipient, please notify the sender immediately by return email and delete this communication and destroy all copies.

1 0

RE: How can I test or simulate an SNMP trap?
by Drew O'Donnell 10 Jul '02

10 Jul '02

Traps are only generated when the volume is full, not the qtree. We parse our logs because messages picks up a qtree error when the quota is exceeded, then we forward this to openview. -----Original Message----- From: Pero, Eric P (ETSD, IT) [mailto:Eric.Pero@thehartford.com] Sent: Wednesday, July 10, 2002 11:06 AM To: 'Eisler, Alex'; 'markallen'; 'Drew O'Donnell'; toasters(a)mathworks.com Subject: RE: How can I test or simulate an SNMP trap? How do you enable the trap for QTREE quotas (or is this already on when you do the "snmp init 1" with the other factory traps)? Eric -----Original Message----- From: Eisler, Alex [mailto:alex.eisler@intel.com] Sent: Wednesday, July 10, 2002 10:30 AM To: 'markallen'; 'Drew O'Donnell'; toasters(a)mathworks.com Subject: RE: How can I test or simulate an SNMP trap? The easier way is just to enable the snmp traps (with 'snmp traps enable' and 'snmp init 1'). This will cause all the factory set traps (like fan failures, disk failures, power supply, temp etc.) to be enabled automatically. (The list is in the MIB, at the end, titled "NetApp trap definitions", on MIB versions 1.5 and up) Alex. -----Original Message----- From: markallen [mailto:markallen@micron.com] Sent: Wednesday, July 10, 2002 16:53 To: 'Drew O'Donnell'; toasters(a)mathworks.com Subject: RE: How can I test or simulate an SNMP trap? Drew, I went into filerview and created a snmp trap for a failed fan. I new what the normal values were and I knew what the abnormal values were. I found out this information by viewing the two files found in /etc/mib. I inverted the values in the SNMP trap by using a normal value to send the trap on. Once it sent the traps and I was finished testing I changed the trip value back to normal so I'm watching for a failed fan or cpu. For example if the normal state of a fan is value=0 and the failed state value=1 I changed the trip value for testing purposes to 0 and then once the testing was over I changed it back to watch for 1. I hope this helps. -Mark -----Original Message----- From: Drew O'Donnell [mailto:Drew@cooperneff.com] Sent: Wednesday, July 10, 2002 7:22 AM To: toasters(a)mathworks.com Subject: RE: How can I test or simulate an SNMP trap? I was interested in this as well, we want to run a scheduled snmp check which will verify our system is working. I would like to write some sort of RSH job to generate a generic trap which we could run weekly. Pulling out a fan is actually how we checked this, but it is not a preferable regular testing procedure. Any Ideas? Drew -----Original Message----- From: Eisler, Alex [mailto:alex.eisler@intel.com] Sent: Wednesday, July 10, 2002 8:21 AM To: 'Mark Allen'; toasters(a)mathworks.com Subject: RE: How can I test or simulate an SNMP trap? Unscrew one of the fans, pull it out, wait for the light to turn orange, push it back in. That should fire 4 snmp traps: 1. FanFail 2. Status critical 3. FanRepaired 4. Status normal Alex. -----Original Message----- From: Mark Allen [mailto:markallen@micron.com] Sent: Tuesday, July 09, 2002 19:29 To: toasters(a)mathworks.com Subject: How can I test or simulate an SNMP trap? All: I set up some snmp traps via filer view and I would like to simulate or test the traps to make sure they works. I set up traps for a failed disk, failed pdu, failed fan, and cluster failover notification. The filer is in production and I would like to find a way to test this out without actually failing the hardware myself. Any ideas? -Mark This communication, including attachments, is for the exclusive use of addressee and may contain proprietary, confidential or privileged information. If you are not the intended recipient, any use, copying, disclosure, dissemination or distribution is strictly prohibited. If you are not the intended recipient, please notify the sender immediately by return email and delete this communication and destroy all copies.

2 1

RE: How can I test or simulate an SNMP trap?
by Pero, Eric P (ETSD, IT) 10 Jul '02

10 Jul '02

How do you enable the trap for QTREE quotas (or is this already on when you do the "snmp init 1" with the other factory traps)? Eric -----Original Message----- From: Eisler, Alex [mailto:alex.eisler@intel.com] Sent: Wednesday, July 10, 2002 10:30 AM To: 'markallen'; 'Drew O'Donnell'; toasters(a)mathworks.com Subject: RE: How can I test or simulate an SNMP trap? The easier way is just to enable the snmp traps (with 'snmp traps enable' and 'snmp init 1'). This will cause all the factory set traps (like fan failures, disk failures, power supply, temp etc.) to be enabled automatically. (The list is in the MIB, at the end, titled "NetApp trap definitions", on MIB versions 1.5 and up) Alex. -----Original Message----- From: markallen [mailto:markallen@micron.com] Sent: Wednesday, July 10, 2002 16:53 To: 'Drew O'Donnell'; toasters(a)mathworks.com Subject: RE: How can I test or simulate an SNMP trap? Drew, I went into filerview and created a snmp trap for a failed fan. I new what the normal values were and I knew what the abnormal values were. I found out this information by viewing the two files found in /etc/mib. I inverted the values in the SNMP trap by using a normal value to send the trap on. Once it sent the traps and I was finished testing I changed the trip value back to normal so I'm watching for a failed fan or cpu. For example if the normal state of a fan is value=0 and the failed state value=1 I changed the trip value for testing purposes to 0 and then once the testing was over I changed it back to watch for 1. I hope this helps. -Mark -----Original Message----- From: Drew O'Donnell [mailto:Drew@cooperneff.com] Sent: Wednesday, July 10, 2002 7:22 AM To: toasters(a)mathworks.com Subject: RE: How can I test or simulate an SNMP trap? I was interested in this as well, we want to run a scheduled snmp check which will verify our system is working. I would like to write some sort of RSH job to generate a generic trap which we could run weekly. Pulling out a fan is actually how we checked this, but it is not a preferable regular testing procedure. Any Ideas? Drew -----Original Message----- From: Eisler, Alex [mailto:alex.eisler@intel.com] Sent: Wednesday, July 10, 2002 8:21 AM To: 'Mark Allen'; toasters(a)mathworks.com Subject: RE: How can I test or simulate an SNMP trap? Unscrew one of the fans, pull it out, wait for the light to turn orange, push it back in. That should fire 4 snmp traps: 1. FanFail 2. Status critical 3. FanRepaired 4. Status normal Alex. -----Original Message----- From: Mark Allen [mailto:markallen@micron.com] Sent: Tuesday, July 09, 2002 19:29 To: toasters(a)mathworks.com Subject: How can I test or simulate an SNMP trap? All: I set up some snmp traps via filer view and I would like to simulate or test the traps to make sure they works. I set up traps for a failed disk, failed pdu, failed fan, and cluster failover notification. The filer is in production and I would like to find a way to test this out without actually failing the hardware myself. Any ideas? -Mark This communication, including attachments, is for the exclusive use of addressee and may contain proprietary, confidential or privileged information. If you are not the intended recipient, any use, copying, disclosure, dissemination or distribution is strictly prohibited. If you are not the intended recipient, please notify the sender immediately by return email and delete this communication and destroy all copies.

1 0

RE: How can I test or simulate an SNMP trap?
by Eisler, Alex 10 Jul '02

10 Jul '02

The easier way is just to enable the snmp traps (with 'snmp traps enable' and 'snmp init 1'). This will cause all the factory set traps (like fan failures, disk failures, power supply, temp etc.) to be enabled automatically. (The list is in the MIB, at the end, titled "NetApp trap definitions", on MIB versions 1.5 and up) Alex. -----Original Message----- From: markallen [mailto:markallen@micron.com] Sent: Wednesday, July 10, 2002 16:53 To: 'Drew O'Donnell'; toasters(a)mathworks.com Subject: RE: How can I test or simulate an SNMP trap? Drew, I went into filerview and created a snmp trap for a failed fan. I new what the normal values were and I knew what the abnormal values were. I found out this information by viewing the two files found in /etc/mib. I inverted the values in the SNMP trap by using a normal value to send the trap on. Once it sent the traps and I was finished testing I changed the trip value back to normal so I'm watching for a failed fan or cpu. For example if the normal state of a fan is value=0 and the failed state value=1 I changed the trip value for testing purposes to 0 and then once the testing was over I changed it back to watch for 1. I hope this helps. -Mark -----Original Message----- From: Drew O'Donnell [mailto:Drew@cooperneff.com] Sent: Wednesday, July 10, 2002 7:22 AM To: toasters(a)mathworks.com Subject: RE: How can I test or simulate an SNMP trap? I was interested in this as well, we want to run a scheduled snmp check which will verify our system is working. I would like to write some sort of RSH job to generate a generic trap which we could run weekly. Pulling out a fan is actually how we checked this, but it is not a preferable regular testing procedure. Any Ideas? Drew -----Original Message----- From: Eisler, Alex [mailto:alex.eisler@intel.com] Sent: Wednesday, July 10, 2002 8:21 AM To: 'Mark Allen'; toasters(a)mathworks.com Subject: RE: How can I test or simulate an SNMP trap? Unscrew one of the fans, pull it out, wait for the light to turn orange, push it back in. That should fire 4 snmp traps: 1. FanFail 2. Status critical 3. FanRepaired 4. Status normal Alex. -----Original Message----- From: Mark Allen [mailto:markallen@micron.com] Sent: Tuesday, July 09, 2002 19:29 To: toasters(a)mathworks.com Subject: How can I test or simulate an SNMP trap? All: I set up some snmp traps via filer view and I would like to simulate or test the traps to make sure they works. I set up traps for a failed disk, failed pdu, failed fan, and cluster failover notification. The filer is in production and I would like to find a way to test this out without actually failing the hardware myself. Any ideas? -Mark

1 0

RE: How can I test or simulate an SNMP trap?
by markallen 10 Jul '02

10 Jul '02

Drew, I went into filerview and created a snmp trap for a failed fan. I new what the normal values were and I knew what the abnormal values were. I found out this information by viewing the two files found in /etc/mib. I inverted the values in the SNMP trap by using a normal value to send the trap on. Once it sent the traps and I was finished testing I changed the trip value back to normal so I'm watching for a failed fan or cpu. For example if the normal state of a fan is value=0 and the failed state value=1 I changed the trip value for testing purposes to 0 and then once the testing was over I changed it back to watch for 1. I hope this helps. -Mark -----Original Message----- From: Drew O'Donnell [mailto:Drew@cooperneff.com] Sent: Wednesday, July 10, 2002 7:22 AM To: toasters(a)mathworks.com Subject: RE: How can I test or simulate an SNMP trap? I was interested in this as well, we want to run a scheduled snmp check which will verify our system is working. I would like to write some sort of RSH job to generate a generic trap which we could run weekly. Pulling out a fan is actually how we checked this, but it is not a preferable regular testing procedure. Any Ideas? Drew -----Original Message----- From: Eisler, Alex [mailto:alex.eisler@intel.com] Sent: Wednesday, July 10, 2002 8:21 AM To: 'Mark Allen'; toasters(a)mathworks.com Subject: RE: How can I test or simulate an SNMP trap? Unscrew one of the fans, pull it out, wait for the light to turn orange, push it back in. That should fire 4 snmp traps: 1. FanFail 2. Status critical 3. FanRepaired 4. Status normal Alex. -----Original Message----- From: Mark Allen [mailto:markallen@micron.com] Sent: Tuesday, July 09, 2002 19:29 To: toasters(a)mathworks.com Subject: How can I test or simulate an SNMP trap? All: I set up some snmp traps via filer view and I would like to simulate or test the traps to make sure they works. I set up traps for a failed disk, failed pdu, failed fan, and cluster failover notification. The filer is in production and I would like to find a way to test this out without actually failing the hardware myself. Any ideas? -Mark

1 0

RE: How can I test or simulate an SNMP trap?
by Eisler, Alex 10 Jul '02

10 Jul '02

You should receive a "STATUS: Normal" message (at least I hope normal is the status on you system) every round hour. The system sends it when at the same time it writes the message to the console. If it doesn't try the following: Snmp traps enable Snmp init 1 This should generate a bunch of messages also if snmp was not initialized. You can also run 'snmp init 0' and then 'snmp init 1' and see what happen. Alex. -----Original Message----- From: Drew O'Donnell [mailto:Drew@cooperneff.com] Sent: Wednesday, July 10, 2002 16:22 To: toasters(a)mathworks.com Subject: RE: How can I test or simulate an SNMP trap? I was interested in this as well, we want to run a scheduled snmp check which will verify our system is working. I would like to write some sort of RSH job to generate a generic trap which we could run weekly. Pulling out a fan is actually how we checked this, but it is not a preferable regular testing procedure. Any Ideas? Drew -----Original Message----- From: Eisler, Alex [mailto:alex.eisler@intel.com] Sent: Wednesday, July 10, 2002 8:21 AM To: 'Mark Allen'; toasters(a)mathworks.com Subject: RE: How can I test or simulate an SNMP trap? Unscrew one of the fans, pull it out, wait for the light to turn orange, push it back in. That should fire 4 snmp traps: 1. FanFail 2. Status critical 3. FanRepaired 4. Status normal Alex. -----Original Message----- From: Mark Allen [mailto:markallen@micron.com] Sent: Tuesday, July 09, 2002 19:29 To: toasters(a)mathworks.com Subject: How can I test or simulate an SNMP trap? All: I set up some snmp traps via filer view and I would like to simulate or test the traps to make sure they works. I set up traps for a failed disk, failed pdu, failed fan, and cluster failover notification. The filer is in production and I would like to find a way to test this out without actually failing the hardware myself. Any ideas? -Mark

1 0

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