Out of Metaspace when redeploying portlets
51 views
Skip to first unread message
Toby Archer
Apr 26, 2018, 11:00:14 AM4/26/18
to uPortal Community
Does anyone else run into this issue:
When I'm working on portlets and I call deployPortletApp frequently, after 10 or so runs I inevitably run into this error. The only thing to do is to kill the tomcat process and restart the portal. On dev this is a minor nuisance, in production it is a big head ache because I'm afraid to update portlets lest I crash the portal.
~TA
2018-04-26 09:41:27 [a8d6b8fa] info [native] ... last message repeated 30 times ...
2018-04-26 09:41:27 [a8d6b8fa] warning [native] Agent not instrumented properly, license cannot be verified
26-Apr-2018 09:41:27.199 WARNING [TransferQueueBundler,uPortal.cacheManager,usd-uportal1-17186] org.jgroups.protocols.TP.sendToSingleMember JGRP000032: usd-uportal1-17186: no physical address for usd-uportal2-30174, dropping message
Exception in thread "https-jsse-nio2-443-exec-241" java.lang.OutOfMemoryError: Metaspace
2018-04-26 09:41:35 [eef7d8f9] info [native] Got ResourceExhausted Notification: Metaspace, not creating Memory Dump
Exception in thread "https-jsse-nio2-443-exec-237" java.lang.OutOfMemoryError: Metaspace
26-Apr-2018 09:44:39.337 WARNING [TransferQueueBundler,uPortal.cacheManager,usd-uportal1-17186] org.jgroups.protocols.TP.sendToSingleMember JGRP000032: usd-uportal1-17186: no physical address for usd-uportal2-30174, dropping message
26-Apr-2018 09:47:51.167 WARNING [TransferQueueBundler,uPortal.cacheManager,usd-uportal1-17186] org.jgroups.protocols.TP.sendToSingleMember JGRP000032: usd-uportal1-17186: no physical address for usd-uportal2-30174, dropping message
26-Apr-2018 09:49:36.551 WARNING [TransferQueueBundler,uPortal.cacheManager,usd-uportal1-17186] org.jgroups.protocols.TP.sendToSingleMember JGRP000032: usd-uportal1-17186: no physical address for usd-uportal2-30174, dropping message
26-Apr-2018 09:49:44.559 WARNING [uP-TaskExec-1067-purgeRawEvents] com.microsoft.sqlserver.jdbc.SQLServerConnection.Prelogin ConnectionID:34429 ClientConnectionId: 71725895-ba22-4f5b-bcd1-e4a7ea9310ca Prelogin error: host usd-sql05.usd.edu port 54648 Error sending prelogin request: Connection reset ClientConnectionId:71725895-ba22-4f5b-bcd1-e4a7ea9310ca
2018-04-26 09:55:02 [fc3c88f9] info [native] ... last message repeated 100 times ...
2018-04-26 09:55:02 [fc3c88f9] info [native] Got ResourceExhausted Notification: Metaspace, not creating Memory Dump
When I'm working on portlets and I call deployPortletApp frequently, after 10 or so runs I inevitably run into this error. The only thing to do is to kill the tomcat process and restart the portal. On dev this is a minor nuisance, in production it is a big head ache because I'm afraid to update portlets lest I crash the portal.
~TA
Andrew Wills
Apr 30, 2018, 12:55:53 PM4/30/18
to Toby Archer, uPortal Community
Toben,
I understand it will “mostly work, a lot of the time,” but I don’t feel that I’m in a position to recommend re-deploying any uPortal web app (incl. portlets) to a running Tomcat under any circumstances, including local dev portals. I always stop, deploy, and start.
Drew
--
You received this message because you are subscribed to the Google Groups "uPortal Community" group.
To unsubscribe from this group and stop receiving emails from it, send an email to uportal-user...@apereo.org.
Visit this group at https://groups.google.com/a/apereo.org/group/uportal-user/.
Toby Archer
May 16, 2018, 9:58:32 AM5/16/18
to uPortal Community, sand...@gmail.com
Darn. That makes pushing changes to production harder since I have to go through the whole chain of command for portal restarts.
Just out of curiosity. I assume that soffits won't have this issue? Or will they still have this issue if they are running in the same tomcat as the portal?
Drew Wills
May 16, 2018, 12:06:03 PM5/16/18
to uporta...@apereo.org
Toben,
On 05/16/2018 06:58 AM, Toby Archer wrote:
> Just out of curiosity. I assume that soffits won't have this issue? Or
> will they still have this issue if they are running in the same tomcat
> as the portal?
Well, it's 100% clear that they *don't* have this issue if they're
running independently of the portal's Tomcat.
But even in the portal's Tomcat, at a minimum I should think you could
use the admin tools to stop & start just the one webapp, and it would
work well.
drew
On 05/16/2018 06:58 AM, Toby Archer wrote:
> Just out of curiosity. I assume that soffits won't have this issue? Or
> will they still have this issue if they are running in the same tomcat
> as the portal?
running independently of the portal's Tomcat.
But even in the portal's Tomcat, at a minimum I should think you could
use the admin tools to stop & start just the one webapp, and it would
work well.
drew
Reply all
Reply to author
Forward
0 new messages